A S P E C T S O F THE GRAMMAR O F
-
I N F I N I T I V E S AND FOR-PHRASES
by
ROBERT A N G E L 0 F A R A C I
S t a t e U n i v e r s i t y of
B.A.,
New Y o r k a t S t o n y B r o o k
1969
SUBMITTED I N P A R T I A L FULFILLMENT
O F THE REQUIREMENTS F O R T H E
DEGREE O F DOCTOR O F
PHILOSOPHY
a t the
MASSACHUSETTS I N S T I T U T E O F
TECHNOLOGY
September,
S i g n a t u r e of A u t h o r .
..
;;
.
\:
;(.
2
.
~
:;:I-
1974
...;-.- r : .....................
D e p a r t m e n t of F o r e i g n L i t e r a t u r e s and L i n g u i s t i c s ,
June 11, 1 9 7 4
C e r t i f i e d by
..... : ~ . ~ J V . ~ ;-J . . : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
T h e s i s Supervisor
A c c e p t e d by
..................
r...............................
Chairman,
Departmental Committee
on G r a d u a t e S t u d e n t s
ASPECTS OF THE GRAMMAR OF
INFINITIVES AND FOR-PHRASES
Robert Angelo Faraci
Submitted t o ~ t h eDepartment of Foreign Literatures and
Linguistics on June 11, 1974, in partial fulfillment of
the requirements for the degree of Doctor of Philosophy.
ABSTRACT
This thesis surveys a number of topics related to
for-phrases. 'We begin by nothe grammar of infinitival ting syntactic and semantic distinctions among three types
of infinitive complements, which are called purpose
clauses, objective clauses, and rationale clauses. It is
shown that these complements must be analyzed as forphrases with sentential objects.
forNext, semantic relations between NP's and phrases are discussed, and the relevance of these relations to the analysis of control phenomena in the three
clause types is considered. It is demonstrated that
"object deletion" in infinitival for-phrases is subject to
conditions on the semantic relations obtaining between the
controller NP and the for-phrase.
Basic semantic differences between gerundives, forphrases and infinitival for-phrases are characterized. Both
are descriptive of motivations for the action,depicted in
the matrix clause. However, gerundive for-phrases characterize motivating factors which are semantically prior
to the action depicted in the matrix clause, while infinitival for-phrases describe intentions which are semantically
posterior to the action characterized by the matrix clause.
In addition, complement subject control in gerundive forphrases is examined.
"Object deletion" in purpose clauses is considered
with respect to the sentential nature of this type of
complement. It is shown that if we analyze the object of
for in this case as a reduced sentence, "object deletion"
in purpose clauses does not violate certain plausible conditions on rules. Following a discussion of tough-p~edicates,
the hypothesis is considered that the complements to the
degree modifiers too and enough are for-phrases with reduced sentential objects.
-
Thesis supervisor: Morris Halle
Title: Professor of Modern Languages
ACKNOWLEDGMENTS
I am especially grateful to my advisor, Morris Halle,
and to the other members of my thesis committee, Noam
-1
Chomsky and Ken Hale, for some very helpful suggestions
for the improvement of an earlier draft of this hhesis.
Noam Chomsky was particularly helpful in pointing out a
number of inadequacies. Morris Halle gave me encouragement
on many occasions when I needed it.
Finally, I want to thank Claire Scarano without whose
moral support and extensive help in preparing this thesis
I would still be a long way from finishing.
CONTENTS
CHAPTER I:
A TYPOLOGY OF SOME INFINITIVE
PHRASES --------------------------------,
7
1.
The distinction between infinitival relative elauses and infinitival purpose
clauses
2.
....................................
Purpose clauses, rationale clauses, and
objective clauses
3.
7
.........................
Relevant aspects of phrase structure
------
26
31
FOOTNOTES TO CHAPTER I
............................
43
CHAPTER 11:
------- for-phrases ----------------
48
SEMAN'TIC RELATIONS AND CONTROL
1.
Parallels with
2.
On semantic relations between NP's and
3.
On control in infinitive for-phrases
4.
Chapter summary
-
------
...........................
48
72
91
APPENDIX A: A DIFFERENT KIND OF NP-FOR-PHRASE
-
APPENDIX B: REMARKS ON EQUI IN PURPOSE
FOOTNOTES TO CHAPTER I1 AND APPENDICES
-----------
106
CHAPTER 111: A DIGRESSION ON GERUNDIVE FORPHRASES
1.
.............................
The prior-posterior distinction
----------
123
123
6
2.
Observations on control in higher-generated
for-phrases ...............................
-
146
.........................
158
FOOTNOTES TO CHAPTER 111
CHAPTER IV:
1.
REMARKS ON OBJECT-DELETION
----------
17 0
"Object deletion" and the internal struc-
------------------
17 0
-------3. Too and enough conStructions ------------FOOTNOTES TO CHAPTER IV ..........................
REFERENCES ......................................
BIOGRAPHICAL NOTE ................................
175
ture of purpose clauses
2.
Remarks on tough-class predicates
187
206
215
217
CHAPTER I
A TYPOLOGY OF SOME INFINITIVE PHRASES
1. The distinction between infinitival relative clauses and
infinitival purpose clauses.
In this section we will describe an ambiguity in sentences like (1) with respect to the underlined infinitive
phrase.
1. Carol bought a rack to hang coats onThe ambiguity of (1) involves the fact that the expression
which refers to or designates the object of purchase is anbiguously construed: It can either be a rack or a rack to
hang coats on.
Otherwise put, the ambiguity of (1) involves
the optional interpretation of the infinitive phrase as a
component of the description of the object which Carol
bought.
As is commonly recognized, the infinitive phrase,
when construed as a component of the referring expression,
functions as a relative clause modifier; and there is, in
fact, an alternative version of sentence (1) on this reading, where the infinitive phrase is introduced by a Whphrase:
2. Carol bought a rack on which to hang coats
Observe that in (2), the infinitive phrase must be con-
8
strued as a component of the referring expression, and the
sentence is unambiguous.
With respect to sentences such as (1) we will say
that the infinitive phrase functions as an infinitival relative clause just in case the infinitive phrase is construed as part of the expression which refers or designates,
and that the infinitive phrase (or the whole sentence) has
an infinitival relative clause reading.
When the infinitive
phrase is not interpreted as part of the referring expression, we will call it a purpose clause (or an infinitival
purpose clause, to distinguish it from constructions which
seem to serve related semantic purposes) and say that the
infinitive phrase or the sentence as a whole has a purpose
clause reading.
That (1) is a genuinely ambiguous sentence can be seen
from the fact that the validity of inferences made from it
depends on whether we take the infinitive phrase to be a relative clause or a purpose clause.
For example, (3) is a
valid inference from (1) only on a purpose clause reading of
(1):
3. Carol intended to hang coats on the rack which she
bought
On the relative clause reading of (I), the inference repre-
sented in (3) cannot be made:
All we know is that Carol
9
baught an object described as a rack to hang coats on; and
we know, in particular, nothing of what she intends to do
with it.
She could, for example, intend to hang her dres-
ses on the rack, and we might add a purpose clause to (1)
to this effect:
4. Carol bought a rack to hang coats on to hang her
dresses on
Observe that, given the purpose clause in (4), (5) is a
pa
valid inEerence from (4):
5. Carol intended to hang dresses on the rack which
she bought
The kind of ambiguity illustrated by examples like (l),
where we have the option of interpreting a certain adjunct
P
phrase as a component of a referring expression, is actually
quite common.
Take, for example, well-known cases like Jane
wrote the letter on the table, where the phrase the letter
I **
I
on the table can be taken as the object of write or on the
table can be taken as a place adverbial modifying the whole
verb phrase write the letter.
There is, however, a further
I--
aspect of the ambiguity of (1) that sets it apart from ex-
~
amples like the one above and engages our interest.
This is
the fact that on the purpose clause reading of (l), the inI
finitival adjunct is used to assert something about the obI
ject a rack.
What is asserted is that the rack serves
the purpose of being something to hang coats on.
This par-
ticular feature is shared by other constructions, as illustrated by the following ambiguous examples:
6. John wrote the story about the Alaskan pipeline
7. Bill took the last picture of his wife
8. Nicola retold the joke about an Italian
(6) can simply be a statement of authorship; i.e., it can
simply mean that John is the author of the story about the
Alaskan pipeline; or it can mean, roughly, that the topic
of the story that John wrote was the Alaskan pipeline.
On
the first reading, the story about the Alaskan pipeline is
a referring expression; i.e., a description of what John
wrote, the object of the verb write.
tl*
On the second reading,
the story is a referrring expression (the object of write),
and it is asserted to be about the Alaskan pipeline.
That
is, the fact that the topic of the story is the Alaskan pipeline is part of what the sentence asserts on this reading.
Similar ambiguities are observable in (7) and (8), whose
readings are indicated by the paraphrases in (9) and (10)
respectively.
9.a. Bill was the photographer of the last picture of
his wife
11
9.b. the last picture Bill took was of his wife
10.a. Nicola was the reteller of the joke about an
Italian
b. Nicola retold the joke as being about an Italian
On the "b-readings" of (7) and (8), given as (9b) and (lob),
the underlined prepositional phrases are not part of any
referring expression.
Further, as in the case of the pur-
pose clause in (I), and the about-phrase in ( 6 ) , these adjuncts are used to assert something about the direct objects
of the sentences: In ( 7 ) , the subject of Bill's last picture is asserted to be his wife; in (8), the subject of the
joke in the version that Nicola told (on Nicola's rendition)
is asserted to be an Italian.
Constructions of this type should be given far more
extensive treatment than we will give them here.
Let me
just add the further observation that there are other adverbial adjuncts which are used to assert something about
either the subject or object of the sentence, adding thus,
a further dimension to the kind of ambiguity we have examined above.
Consider, for instance, examples of the fol-
lowing type:
11. El Greco painted the Cardinal without his glasses
12. The clown put the tie on upside down
13. Alexander recited the poem without the usual
soppiness
'
12
In (ll), it could either be the Cardinal or El Greco who is
without his glasses; in (12), it could be either the clown
or the tie that i~ upside down; in (13), it could be either
Alexander or the poem itself that is without the usuaJ soppiness.
Thus, there seems to be a general problem of ad-
juncts with apparent adverbial force which attribute some
property or properties to a noun phrase in the containing
sentence.
The phenomenon of purposei&9ausesLis:a particu-
lar instance of this general problem.
Our task now is to demonstrate that the ambiguity of
(1) (and of (6-8)) is structural.
Specifically, we wish
to show that the infinitive phrase as an infinitival relative forms a syntactic constituent with the nominal a rack
to its left, while as a purpose clause it does not form a
constituent with a rack.
To bring out the structural nature
of the ambiguity, we can rephrase the ambiguity as one in
'
the construal of the object noun phrase of the verb buy:
Is
the object a rack or a rack to hang coats on?
(The object
NP of buy is, of course, a referring expression:)
Consider,
for example, the two possible passive versions of (1):
14. a rack to hang coats on was bought by Carol
15.
a' rack
was bought by Carol to hang coats on
In (14), where the infinitive phrase is treated by the Passive Pule as a constituent of the object NP, only a relative
clause reading is possible; the infinitive phrase is interpretable only as a component of a referring expression.
(15)
on the other hand, has its infinitival adjunct unaffected by
the Passive rule; it has not been taken as a constituent of
the object NP which is preposed.
In this situation, a rack
is the referring expression (the object NP) and the infinitival adjunct is interpreted as a purpose clause.
It is
possible that (15) is actually ambiguous, having a relative
clause reading as well.
We can account for this possibility
by deriving (15) on the relative clause reading from (14)
by the well-known rule of Extraposition from NP.
Thus, we
(1.53, allowing the infinihave two possible derivations f o ~
tive phrase to be interpreted either as a relative clause
or a purpose clause.
Consider next the following pseudo-cleft examples.
16. what Carol bought was a rack to hang coats on
17. what Carol bought to hang coats on was a rack
(16) is quite straightforward: the infinitive phrase is part
of the postcopular focus constituent and is interpretable
only as an infinitival relative clause.
On the other hand,
the infinitive phrase in (17) is unambiguously interpreted
as a purpose clause.
Notice that if we substituteFfor the
14
infinitive phrase in (17) its "Wh-relative clause version",
the sentence is ungrammatical:
18. "what Carol bought on which to hang coats was a rack
However one wishes to derive pseudo-cleft sentences, it is
a fact about them that what appears in focus position cannot
be an immediate constituent of a complex noun phrase.
In the
instance of (17) and (18), the relevant derivative observation is that the heads of relative clause contructions cannot appear in focus position without the relative clause.
Thus, in a sentence like
19. what John ate that Sarah made was the cookies
that Sarah made cannot be interpreted as a relative clause
modifier of the cookies.
In the familiar way.. in which pseu-
do-cleft sentences are described, we would say that (19)
does not "correspond to" the sentence, John ate the cookies
that Sarah made (it is not even clear;that,(l9) is grammatical).
One of the pseudo-cleft sentences which would rightly
be said to "correspond to" this latter sentence would be,
what John ate were the cookies that Sarah made.
(For a ree
cent, interesting discussion of the pseudo-cleft construction,
see Higgins 1973.
Higgins rejects the standard transforma-
tional analyses for these sentences in favor of an analysis
which virtually identifies the deep structures of pseudo-,,
clefts with their surface structures.)
From these observations, it follows (correctly) that
(17) does not "correspond to" (1) on the relative clause
reading of the latter.
a
According to our observations, -
rack and to hang coats on do not form a complex noun phrase
at any stage in the derivation of (17).
Like our observa-
tions about the Passive, these facts indicate that the two
readings for (1) are associated with different structural
descriptions for the sentence.
On the relative clause read-
ing, a rack and to hang coats on form a single constituent;
on the purpose clause reading, they do not.
Now observe,
incidentally, that the infinitive phrase in (I), interpreted
as a purpose clause, can itself appear in focus position in
a pseudo-cleft sentence:
20. what Carol bought a rack for was to hang coats on
(The for which shows up in (20) will be dealt with later on.)
The impossibility of a relative clause interpretation for
the focus constituent~iof (20) can be seen when we substitute
the "Wh-version" for it:
21.
* what
Carol bought a rack (for) was on which to hang
coats
TJe
see, then, that pseudo-cleft constructions point up cku-
cia1 syntactic differences between infinitival relAtives and
infinitival purpose clauses.
Let us pause to note that quite similar observations
"-
can be made for (6-8). Consider, for example, (7). The
two possible passive versions are as follows:
22. the last picture of his wife was taken by Bill
23. the last picture was taken by Bill of his wife
Each of these unambiguous examples has one of the possible
readings ~ 6 f( 7 ) .
(22) has the reading corresponding to (9a);
.
(23) has the reading corresponding to (9b)
In (22), of his
wife is interpreted as part of the referring expression and
is analyzed syntactically as a constituent of the surface
subject NP which has been preposed by;the Passive rule from
-
the position of the direct object of take in (7). In (23),
corresponding to (9b), of his wife does not form part of a
referring expression and is unaffected by the Passive rule.
In this case, the object of take is simply the last picture.
(Incidentally, (23) does not seem to me to have an alternative reading corresponding to the reading of (22), where the
of-phrase at the end of the sentence could be taken as an
extraposed complement.
Compare the ambiguous (15). The
conditions under which an extraposed complement reading is
in general available are, in any event, quite complicated and
poorly understood.)
Further, consider the following pseudo-cleft examples:
17
24. what Bill took was the last picture of his wife
25. the one that Bill took the last picture of was
his wife
(24),
of course, has only the (9a) reading; the material to
the right of the copula must be analyzed as a constituent,
and this focus constituent forms a referring expression.
situation is different for (25).
The
Here, the of-phrase cannot
be interpreted as the complement of picture, since there are
severe restrictions on pseudo-clefting out of a complex noun
phrase.
(This observation can easily be translated in terms
of your favorite theory of pseudo-clefts.)
Thus, aside from
the relative clause examples we discussed above, if we take
a sentence like
26. John is married to the woman next to Bill
we find that it is impossible to have Bill in focus position
in a "corresponding" pseudo-cleft sentence:
27. *the one who John is married to the woman next to
is Bill
Also, reconsidering the ambiguity of the example
28. Jane wrote the letter on the table
where on the table
can be understood either as a place ad-
verbial modifying the verb phrase or as part of the referring
expression the letter on the table, notice that only the
.
place adverbial reading for the on-phrase is possible in the
following pseudo-cleft:
on was the table
29. what Jane wrote the letter The impossibility of pseudo-clefting the object of a PP
nominal complement insures the non-ambiguity of (29).
When the adjuncts are not construed as components of
referring expressions (the infinitive phrase of (1) on the
purpose clause reading; the prepositional phrases of (6-8)
on their "b-readings"), the NP's to their left (the direct
objects) can be replaced by pronouns, since they are understood to have reference by themselves, without the adjuncts.
Observe the non-ambiguity of the following cases:
30. Carol bought it to hang coats on
(cp. (1))
31. John wrote it about the Alaskan pipeline
32. Bill took it of his wife
(cp.(6))
(cp. ( 7 ) )
33. Nicola retold it about an Italian (cp. (8))
The fact that it cannot form a constituent with the ad-junct
to its right is quite clear.
If we try, for example, to
force a relative clause reading for the infinitive phrase in
(30) by substitution with the "Wh-version", the result is
ungrammatical.
34, *Carol bought it on which to hang coats
If we try to treat-.;ity'adjunct
- .as a constituent with respect
to Passive or the demonstrative that
+
adjunct as the focus
constituent in a pseudo-cleft sentence, the result is also
ungrammatical.
pronoun
(We use the demonstrative that because the
not a
focus constituent.
Cf. Carol
bought that, what Carol bought was that; Carol bought it,
*what Carol bouaht was it.)
35. *it to hang coats on was bought by Carol
36. *what Carol bought was that to hang coats on
37. *it about the Alaskan pipeline was written by John
38. *what John wrote was that about the Alaskan pipeline,
etc.
I n semantic terms, the non-ambiguity of (30-33) makes sense:
If the adjunct is interpreted as a nominal modifier, the head
of the construction does not have reference by itself; rather,
the whole expression, head
+ modifier, is what refers.
are interpreted as having referDefinite pronouns, which ence, consequently cannot substitute for the heads of modified nominal constructions.
Thus, the adjuncts in sentences
(30-33) do not form referring expressions with the nominal
it to theiL left. Moreover, as examples (35-38) demonstrate,
the adjuncts do not form syntactic constituents with the
- ,
20
pronoun it, and this contrasts with the fact that when the
adjunct is a component of the referring expression, the
whole referring expresssion is treated syntactically as a
constituent.
The ambiguity of interpretation for (1) and
(6-8), then, which revolves around the question whether the
nbminal to the left of the adjunct has reference, correlates
with an ambiguity of structure. Notice, incidentally, that
the point of the discussion could have been made just as well
if, instead of pronouns, we used NP's with deictic determiners (this/that rack) or with possessive determiners (his/Mary's rack) , etc. 1
Finally, the fakt:that purpose clauses are not syntactically noun phrase complements(and, in particular, not relative clauses) is shown by the fact that it is possible to
"chop" NP's from them.
In order to demonstrate this, how-
ever, it is first necessary to digress briefly.
So far in
this discussion, we have examined purpose clauses where the
deleted NP is an object (a direct object or the object of a
preposition), but an interesting property of purpose clausesone which we will consider much more extensively at a later
point- is that this deletion operation is not restricted to
object NP's.
Subject NP's can also delete.
the following examples:
Thus, contrast
39. a. we bought this dogi for our children to play
with to play with our chil-
b. we bought this dogi -i
dren
40. a. John rented the airplanei to take i
Mongolia
b. John rented the airplanei i
Mongolia
to
to take him to
In this respect, then,ipurpose clauses resemble infinitival
relative clauses: the NP missing in surface structure can
be a subject, object, or object of a PP.
Consider now the following example:
41. Wolfgang bought this violin
j
to play sonatas on
-1
Suppose we relativize the NP sonatas in (41). What we get
is an ill-formed noun phrase:
42. *...the sonatasi which Wolfgang bought this violin to play - on
j
-1
However, the ill-formedness of (42) does not (necessarily) reflect the impossibility of relativizing out of a purpose
clause.
The fact is that once an NP has been chopped out of
a VP, no other NP can be chopped out of that VP.
I believe
this generalization was first observed by Bruce Fraser in
connection with Tough-Movement examples.
Thus, consider the
sentence,
43.
it is easy to play sonatas on this violin
Now, either sonatas or this violin can be moved to subject
position to produce:
44. sonatasi are easy to play i
on this violin
45. this violini is easy to play sonatas on But once Tough-Movement (or its equivalent on some other
theory) has applied either as in (44) or (45), the remaining NP in the VP complement to the tough--typepredicate
cannot be chopped, as illustrated by the contrast of the
following examples.
46. a. the sonatasi which it is easy to play -i
on this violin
b. the violin
on
i
which it is easy to play sonatas
-i
47. *th&;violin which sonatas are easy to play j
i
48. *the.sonatas
j
which this violini is easy to
I am not aware that any explanation of the ill-formedness
of examples like (47) and (48) has been proposed2, but it
seems plausible to conjecture that some difficulties in the
23
application of perceptual stragegies are involved here.
But,
whatever the explanation is, these facts indicate the difficulty of testing relativizability out of purpose clauses with
examples like (42) and motivate the introduction, at this
point in our discussion, of the "deletability" of purpose
clause subjects (cf. the b-examples of (39) and (40)), since
this property allows us to bypass the difficulty.
Thus, notice that children in (39b) and MongoJia in
(40b) can relativize successfully:
49. the children that we bought this dogi
j
play with
-i
to
-1
50. the place
that John rented the airplanei
j
to take him to
-j
-i
Also, notice the grarnrnaticality of these cleft examples. > :
(Clefting is another standard illustration of "chopping"
phenomena.)
51. it was our children that we bought this dog
j
ito play with j
52. it was Mongolia that John rented the planei
j
i
to take him to
-j
i
Therefore, because of the grammaticality of (P9-52), we
can conclude that the ill-formedness of (42) is due not to
any prohibition against relativizing out of purpose clauses,
24
but rather to certain general conditions on chopping rules.
The fact that purpose clauses can be "chopped" out of is
clear evidence that they are syntactically distinct from
relative clauses, since chopping out of relative clauses is
a prime example of the violation of the Complex Noun Phrase
Constraint (Ross dissertation, 1967), or its equivalent on
some other theory (cf. e.g., Chomsky's (1973) recent discussion of adjacency.)
On the basis of the evidence considered in the foregoing discussion, we conclude that the structural ambiguity of (1) is to be represented approximately as follows.
(These deep structure proposals will be refined in the course
of this thesis.
Further, I do not wish to defend any parti-
cular claims for the deep structure of relative clause constructions, and represent the relative clause merely as a
complement to a nominal.
way of doing things. )
This seems to me to be a neutral
(53) is the structure of (1) on the
relative clause reading; (54) the structure on the purpose
clause reading.
In subsequent sectiolns,we will be concerned with a more
detailed description of purpose clauses, and remarks about
infinitival relatives will be limited principally to whatever concerns their relation to purpose clauses.
I should
add that the kind of purpose clause under discussion, as distingGished from infinitival relatives, has been recognized
in traditional grammar.
For example, Jespersen (Modern- -
Erlglish Grammar, Volume V, 15.2-15.3; 16.49) discusses these
clauses and refers to them as retroactive, meaning that they
have passive sense without passive form; i.e., they allow
"object deletion".
He does not, sorfar as I know, consider
subject dgletion in purpose clauses.
2. Purpose clauses, rationale clauses, and objective clauses.
In the last section, we isolated a type of infinitive
cemplement which we called a purpose clause and discussed
some of the ways in which it could be distinguished from an
infinitival relative.
Here, we extend our observations to
distinguish between purpose clauses and what we will call
rationale clauses and 6bjective clauses.
The discussion
will be exploratory; we will be considering mainly facts
which we will need to understand before we can deepen
analysis at later points.
our
Later on, we will attempt to sys-
tematize our observations and discuss the theoretically interesting aspects of the topic.
Consider the following example:
1. Bill bought the pianoi for Mary to practice music
2. Bill bought the pianoi for Mary to practice music
In (I), we have a purpose clause with the deleted object of
on anaphorically related to the piano, the object of buy.
In (2), there is no deletion, and the sentence is word-forword identical with (l), except that rather than a "deletion
site" after on, we have a pronoun object it, which is
anaphorically related to the.,piano,the object of buy.
The infinitive phrases in these sentences serve distinct semantic functions: In line with our previous observations about purpose clauses, the purpose clause in (1)
designates the purpose or function served by the piano; the
piano serves as something for Mary to practice music on.
Moreover, we should add that the purpose clause designates
Bill's intentions for the piano.
That is, !we understand
from (1) that it is Bill who decides that the purpose of the
piano is going to be what the purpose clause says it is. CRecall the discussion of sentence (3) of the last section;)
In
(2), the infinitive phrase defines the reason for Bill's ac-
tion; it answers the question, why did Bill buy the piana?, in
the motivational sense.
(2) would be appropriate, for exam-
ple, to a situation in which Mary has refused to practice on
a certain piano unless it were bought by Bill, so that Bill's
purchase of the piano is motivated by his desire to have
her practice on it.
Infinitive phrases of the type occurring in (2) have
been called traditionally result clauses. (Cf. Jespersen,
MEG, Volume V, 16.54-16.6.)
The matrix sentence defines a
condition for obtaining the result specified in the infinitive phrase.
In this discussion, we will use the term
rationale clause.- rather than result clause, for reasons which
will be clearer after further discussion.
The term
rationale clause is justifiable, since such a clause designates the motivation for the action depicted in the matrix
clause.
In the case of kationale clauses, there is no require*
ment of coreferentiality between an NP in the rationale
a
clause and an NP in the matrix clause.
3. Bill bought the piano for Mary to gain a fuller appreciation of keyboard music
On the other hand, deletion of an NP in a purpose clause under identity with an NP in the matrix clause is obligatory.
A further difference between rationale clauses and pur-
pose clauses is that rationale clauses can be introduced
by in order (for X ) to, while purpose clauses cannot:
4. *Bill bought the pianoi in order for Mary to practice on
Bill bought the piano
order for Mary to practice
6. Bill bought the piano in order for Mary to gain a
=
fuller appreciation 6f keyboard music
Other differences will be examined later on.
Closely related to rationale clauses are what I will
call objective clauses.
lowing examples:
Consider the ambiguity of the fol-
7. the teacher sent the student to the office to annoy
the principal
8. John trains the new recruits to make a living
9. Ned hired the lawyer to protect his son
In (7-9) there is an ambiguity.
The subject of the infini-
tive phrase can be either the subject or object of the matrix sentence; i.e., the subject of annoy in (7) can be either the teacher or the student,- and so on.
r
On the inter-
pretation where the subject of the infinitive phrase is the
subject of the matrix sentence, the infinitive-phrases ar;e
rationale clauses; on the other interpretation they are what
I call objective clauses.
The difference between rationale clauses and objective
clauses is perhaps clearer from the following examples, in
which the use of reflexive pronouns makes clearer the possibilities {in the interpretation of the subjects of the infinitive complements.:
10. a. John trains the new recruits to make a living for
himself (rationale)
b. John trains the new recruits to make a living for
themselves (objective)
11. a. Ann sent Alex to Toronto to spend some time by
herself
(rationale)
1
b.
(884'b NIL I ,
I,,'/,,
18
8,
t.,
8,
8 7 ,
1
,
3
,IN
,,8
Ann s e n t Alex t o Toronto t o spend some t i m e by
Himself
(objective)
N o t i c e t h a t o b j e c t i v e c l a u s e s ( l i k e purpose c l a u s e s ) cann o t be i n t r o d u c e d by i n o r d e r ( f o r X ) t o .
Thus, compare ( 1 0 )
and (11) w i t h (123: and ( 1 3 ) below:
1 2 . a . John t r a i n s t h e new r e c r u i t s i n o r d e r Co make
a l i v i n g f o r himself
b.*John t r a i n s t h e new r e c r u i t s i n o r d e r t o make a
l i v i n g f o r themselves
1 3 . a . Ann s e n t Alex t o Toronto i n o r d e r t o spend some
t i m e by h e r s e l f
-b.*Ann s e n t Alex t o Toronto i n o r d e r t o spend some
t i m e by h i m s e l f
N o t i c e a l s o how i n t r o d u c t i o n of t h e p h r a s e i n o r d e r t o d i s ambiguates s e n t e n c e s ( 7 - 9 )
i n f a v o r of t h e r a t i o n a l e c l a u s e
interpretation:
14.
t h e teacher sent the student t o the o f f i c e i n
o r d e r t o annoy t h e p r i n c i p a l
15.
John t r a i n s t h e new r e c r u i t s i n o r d e r t o make
a living
16.
Ned h i r e d t h e lawyer i n o r d e r t o p r o t e c t h i s son
3. Relevant aspects of phrase structure.
We argue in this section that rationale clauses have
a different deep structure source from purpose clauses and
objective clauses.
The thesis here is that purpose clauses
and objective clauses are generated as complements inside
the VP, while rationale cdauses are generated as daughter
constituents of PredP (or E, or however one wishes to label
the node) -- i.e., outside the VP.
The first argument for our contentions has to do with
preposability.
Chomsky (1965: pp. 102ff.) argued that only
phrases outside the VP could be preposed to the beginning
of the sentence.
The preposing of constituents originating
in the VP is acceptable only if the resulting sentence is
given some sort of topicalization intonation.
These ideas
have been somewhat sharpened and elaborated in a very interesting paper by Williams (1971).
Rationale clauses distinguish themselves from purpose
clauses and objective clauses by their preposability.
Con-
sider the following objective-rationale clause pairs:
1. a. Maryi hired John to protect heri (objective)
b. Mary hired John to protect hersezf (zationale)
2. a. Ann sent Ned to NY to spend some time by himself
(objective)
b. Ann sent Ned to NY to spend some time by herself
(rationale)
3. a. Sam trains the new recruits to make a living for
themselves
(objective)
b. Sam trains the new recruits to make a living for
himself
(rationale)
Observe that preposing the infinitive phrases is possible
only in the case of rationale clauses:
4. a. *to protect heri, Mary i hired John
b. to protect herself, Mary hired John
5. a. *to spend some time by himself, Ann sent Ned to NY
b. to spend some time by herself, Ann sent Ned to NY
6. a. *to make a living for themselves, Sam trains the
new recruits
b. to make a living for himself, Sam trains the new
recruits
Notice, also, that purpose clauses are not preposable.
7. a. John bought the piano for Mary to practice on it
(rationale)
b. John bought the piano for Mary to practice on
(purpose
8. a. for Mary to practice on the piano, John bought it
b. *for Mary to practice on, John bought the piano
The second argument has to do with what we might term the
degrees of "dependency" of the clause on the matrix verb.
For
any of the infinitive phrases we are dealing with, it seems
incorrect to describe this dependency in terms of syntactic
subcategorization.
Rather, it seems that the relatiohship
is better described in terms of semantic compatibility.
Each clause type appears to have some inherent semantic function which imposes conditions on the types of matrix predicates with which it is compatible.
In the spirit of the
analysis developed in Williams (1971), I will show that the
compatibility conditions are more restrictive in the case of
bbjective clauses and purpose clauses than in the case of
rationale clauses.
The reasoning here is that the tighter
the dependency of the clause on the verb, the lower its level of "embedding. (See Williams 1971:
10, where he dis-
cusses the degree to which an item can subcategorize a verb.)
In the case of rationale clauses, there are items other
than the matrix predicate on which such a clause can be dependent.
Basically, rationale clauses are compatible with vo-
suflitional predicates, conditional predicates (necessary, ficient, need), and with some of the modals.
If a modal or
conditional predicate is present, the volitionality of the
predicate-which follows it is irrelevant.
These points are
illustrated in the examples below:
9. John ("accidentally) let the cats out of the room
(in order) to have some peace and quiet
34
10. Max slept eight full hours last night (in order) to
make sure he would be alert this morning, "although
he had intended only to doze off for a few minutes
11. *Ivan was tall (in order) to attract attention
12. Ivan must2-be tall (in order) to attract attention
13. Ivan needs to be tall (in order) to attract attention
14. it is sufficient for Ivan to be tall (in order) to
attract attention
In ( 9 ) , we are dealing with a willful action in the matrix
clause, as we can see from the anomaly that results with the
addition of the adverb accidentally.
For (10) to be accept-
able, we have to be able to infer that Max's sleeping eight
hours was intentional. The point about the volitionality
of the matrix predicate is that in sentences like (9) and (10)
the rationale clause indicates the intent of the matrix action.
As a result, the although-clause makes (10) anoma-
lous, since it contradicts the implication of intentionality.
In (ll), where Ivan can have no control over his height.+and,
hence, where there can be no possibility of intent, the
presence of a rationale clause results in an anomaly.
But
notice how in (12-14), with the addition of the modal must
or the conditional predicates need, be sufficient, which
serve to specify the predicate be tall as a condition on a
result, the presence of a result clause is perfectly accept-
'
able.
(Recall the traditional term, result clause.)
Thus, when we attempt to identify the conditions which
allow for rationale clauses, we find that they depend on
the nature of the matrix predicate only in a ,!broad sense.
In no way can we say that verbs select for rationale clauses.
Turning to purpose clauses, we find that the -situation
is rather different.
For example, purpose clauses are com-
patible with some, but not all volitional predicates.
Thus,
we may have sentences like
.L.
15.
Mary built the board to play chess on
16.
Mary bought the board to play chess on
17.
Harold made the stove to cook his meals in
18.
Harold used the stove to cook his meals in
but not
19. *Mary destroyed the board to play chess on
20. *Mary repaired the board to play chess on
21. *Harold painted.- the stove to cook his meals in
22. *Harold cleaned the stove to cook his meals in4
In general, purpose clauses are compatible with certain
fairly broad classes of predicates in English.
Among
them
are (1) predicates of transaction, such as give,buy,sell,
take,
-
steal, borrow, lend, (2) transitive verbs of motion,
such as send, bring, take,
(3) verbs of creation, such as
build, construct, devise, make, and (4) the verb use.
The
conditions of compatibility with the main predicate are
much more ~estrictivein the case of purpose clauses than in
the case of rationale clauses.
What I have termed objective
clauses are usually complements to verbs of motion like send,
-
bring, and take.
An objective clause characterizes an objec-
tive which involves the passive or active participation of
the individual or thing which is acted on in the matrix :clause as in they sent him to the mountain to be crucified,
they sent her to;the mountain to crucify him, the samples
were taken to the hospital to be examined, the police were
brought in to oversee the demonstration.
We conclude, then, that for purpose and objective
clauses, there is a much closer semantic association with
the matrix predicate than for rationale clauses.
A third argument for the thesis of this section has to
do with relative order.
Again, we draw on the results of
Williams (1971). When both a purpose or objective clause
and a rationale clause figure in the same sentence, the purpose clause or objective clause must always precede the
rationale clause ,(discounting, of course, the possibility of
preposing the rationale clause).
23. a.
Marc bought Fido to play with -(in order) to
please Anita
ale clause)
(Purpose clause precedes ration-
37
23. b. *Marc bought Fido (in order) to please Anita to
play with
(Rationale clause precedes purpose
clause)
24. a.
Ben took Alice to Bston to amuse herself to
please himself
(Objective clause precedes
rationale clause)
b. *Ben took Alice to Boston to please himself to
amuse herself
(Rationale clause precedes
objective clause)
The relative order of the infinitive phrases is explained if
the sources for purpose clauses, objective clauses, and ray
tionale clauses are as we contend.
If rationale clauses are
generated as immediate constituents of PredP to the right of
VP, then they will always be generated to the right of purpose clauses or objective clauses, which are generated as
constituents of VP.
A fourth argument concerns certain aspects of the control
properties of the infinitive phrases under discussion.
In
considering Equi-NP deletion, Williams (1971: 13) puts forth
"the deleting (or 'control') NP cannot be
the hyp~thesis~that
lower in the matrix tree than the clause containing the NP
to be deleted. 11 5
In the light of this plausible hypothesis, let's consider the following observations.
Clearly, we have seen that
the subject of an objective clause can be controlled by the
38
object of the matrix verb, while the subject of a rationale
clause must be controlled by the matrix subject.
25.
Mort sent his robot to us to get the prize
In ( 2 5 ) , when the subject of the infinitive complement is
controlled by the robot (the object of send), the sentence
has an objective clause reading; when the complement subject
is controlled by Mbrt (the matrix subject), the sentence has
a rationale clause reading.
Since the matrix object is gen-
erated lower in the matrix tree than the matrix subject, and
since the matrix object cannot delete the complement subject
of a rationale clause, we conclude that the objective clause
is generated lower in the matrix tree than the rationale
clause.
Furthermore, since, according to Williams's cri-
teria, the direct object is generated in the lowest level of
the matrix tree, it follows from the hypothesis that the objective clause must also be generated in the lowest level of
the matrix tree. 6
For purpose clauses, one NP in the infinitive phrase is
controllable by the direct object of the matrix predicate,
indicating that they, too, are constituents of the lowest
VP-level of the matrix tree.
Furthermore, consider the fol-
lowing contrast between purpose clauses and rationale
clauses.
For purpose clauses, at least two control rules are
needed: one which effects the deletion of an NP in the infin-
39
itive phrase under identity with the NP with which the purpose clause is "associated" (i.e., that NP to which the> ;
purpose or function is ascribed) and a rule of Equi.
Both
rules are operative in the example below.
26. a.
b.
Maryi bought a rag doll.
J--i
when she had time off
to play with
-j
to play with it
Maryi bought a rag doll
j -i
j
when she had time off
In (26a), with a purpose clause, we have two deletions as
indicated.
The deletion of the complement subject by the ma-
trix subject Mary is effected by the Equi rule referred to
above.
In (26b), we have a rationale clause, and a similar
Equi process deleting the complement subject under identity
with the matrix subject Mary.
However, the difference be-
tween the Equi rule in (26a) and the Equi rule in (26b) is
shown when we add indirect objects to the verb buy:
27. a.
Mary bought her daughteri a rag doll
j -i
to play with
when she had time off
-j
b.
Mary i bought her daughter a rag doll
j -i
to play with it when she had time off
j
In the case of the purpose clause, the introduction of an indirect object changes the controller of the complement subject; in the case of the rationale clause, the control of the
40
complement subject remains unaffected by the introduction of
an indirect object.
Since the indirect object controls the
complement subject in (27a) but not in (27b), it suggests
that (1) the purpose clause is generated at the lowest VPlevel, since the indirect object is presumably generated at
that level. (The fact that the direct object also deletes
into the purpose clause points to the same conclusion.); and
(2) the fact that the indirect object cannot delete the subject of a rationale clause could indicate that the rationale
clause is generated at a higher level in the matrix tree,
making its subject "inaccessible" to the indirect object of
the matrix ~ e r b .
The final argument that we will give here is that
rationale clauses cannot form part of a verb phrase in focus
position in pseudo-cleft sentences, while objective clauses
and purpose clauses can.
28. a. Sam read The Master and Margarita to amuse himself
b. *What Sam did was read The Master and Margarita
to amuse himself
29. a. Alice played hookey to anger her parents
b. *What Alice did was play hookey to anger her parents
The b-examples of (28) and (29) demonstrate that rationale
clauses cannot form part of the focus constituent. Contrast
these paradigms with the following:
30. a. Ben brought Alice home to amuse herself
b. What Ben did was bring Alice home to amuse herself
31. a. Marc bought Fido to play with
b. What Marc did was buy Fido to play with
The b-examples of (30) and (31) are grammatical, in contrast
with those of (28) and (29). That is, the focal verb phrase
constituents in the b-examples of (30) and (31) include respectively the objective clause and the purpose clause as subconstituents.
On the basis of the evidence considered in this section,
we conclude that rationale clauses are generated on a higher
level on the matrix tree than objective clauses and purpose
clauses. Assuming that something resembling the following
phrase-structure (PS) rules are correct for English,
we hypothesize that rationale clauses are introduced by (32),
expanding
(=Pred P) , while objective clauses and purpose
clauses are introduced by (33), expanding VP. In tree-diagram
form, the structure is as follows.
v
NP
'S ( o b j e c t i v e c l a u s e s ;
p u r p o s e clauses)
FOOTNOTES
CHAPTER I
1.
It has been suggested to me by various people that pur-
pose clauses as in (1) must really be treated as a special
type of relative clause.
The basis for this is the striking
similarities between purpose clauses and relative clauses
with respect to matters of control.
But why assume that
such control properties are restricted to relative clause
constructions? As we have noted, in other crucial respects,
purpose clauses and infinitival relatives differ vastly:
Purpose clauses do not form constituents with the nominal
that binds them; the nominal head of a relative clause construction does not refer by itself; etc.
Thus, purpose
clauses do not have the-syntactic properties and semantic
functions in general that define relative clauses,
The
point about similarities in control properties will be
taken up at length.:
2.
It has recently come to my attention that Ivonne
Bordelois proposes an explanation for such data in her forthcoming dissertation.
I am not familiar with the details
of her proposal, but I know that she judges sentences like
(47) to be more grammatical than sentences like (48).
I
agree with her judgments, although I do not find sentences
like (47) totally acceptable.
3.
The relationship between volitionality and intention is
more complicated than we have implied in the text.
For
example, consider the following question,
i, what did John say to make Mary so angry?
In this example, volitionality and intention are independent,
for even though the act of saying something is volitional,
it is not necessary that Mary's resultant anger represent
any intent of John in saying what he did.
Thus, in sen-
tences like (i), where there'.isno implication of intention (but, rather, of the actualization of a result), we
get phrases like so or as he did and the like:
ii. what did John say to make Mary as angry as he did?
iii. what did John say to make Mary as angry as she is?
Note also that the same kind of interpretation of the infinitive phrase is possible withz- the modal must in its epistemic sense:
iv. John must have said some awful things to make Mary
as angry as he did/ as she is
There are, in addition, other cases where the implication of
an actualized result is present.
45
v. John overslept yesterday, to make this the third
day in a row that he hasn't shown up for work on time
Two things are to be noted immediately about such sentences:
(1) the infinitive phrase cannot prepose:
vi. to make this the third day in a row he hasn't shown
up on time for work, John overslept yesterday
For (vi) to be acceptable, the preposed infinitive phrase
must be understood as a designation of John's intention.
(2) on the non-intention reading of (v), the subject of the
infinitive phrase is understood not as John but as John's
having overslept yesterday or something of this sort.
4.
On the basis of such data, we could show that the dis-
tribution of purpose clauses and infinitival rklatives differ, though they overlap.
1.
E.g., in a sentence like
Mary used a pan to fry eggs in
the infinitive phrase can be interpreted either as a purpose clause or as an infinitival relative.
But if we change
the matrix predicate as in:
ii.
Mary displayed a pan to fry eggs in
we eliminate the purpose clause reading.
pan cannot be replace by a pronoun:
-
Thus, in (ii), a
iii. *Mary displayed it to fry eggs in
Also, compare the array of pseudo-cleft sentences "corresponding to" (i) with the array of those "corresponding to"
.
(iii)
iv.a.what Mary used was a pan to fry eggs in
b.what Mary used to fry eggs in was a pan
c.what Mary used a pan for was to fry eggs in
v.a.what Mary dispJayed was a pan to fry eggs in
b.*what Mary displayed to fry eggs in was a pan
c.*what Mary displayed a pan for was to fry eggs in
5.
Williams notes a class of exceptions to this putative
generalization, as in the sentence
i. Bill hit John for stealing grapes (WilLiams's (54))
where, by cerkain of his criteria, the understood subject
John of
-
stealing grapes is generated lower in the matrix
tree than the for-phrase complement.
It is interesting to
note that if we add the modal must to (i), the subject of
stealing grapes becomes ambiguous.
ii. Bill must hit John for stealing grapes
On one reading of (ii), John's punishment for stealing
grapes is being hit by Bill; on the other reading, Bill's
punishment for stealing grapes is hitting John.
See Chapter
I11 and Faraci (1971) for further discussion.
6.
This may point up a problem in Williams's analysis, since
the first or lowest level in the tree is reserved for the
verb and those items for which it iS strictly subcategorized.
CHAPTER I1
SEMANTIC RELATIONS AND CONTROL.
In this chapter, we will elaborate on the analysis of
the clause types discussed in Chapter I. In the first section,
I will argue that purpose, objective and rationale clauses
for-phrases; they are to be generated
are to be analyzed as in deep structure as objects of the preposition -for. In the
second section, I will introduce certain dkstinctions in
semantic relations that exist between NP's and for-phrases,
and I will show, in the third section, the relevance of these
distinctions to control phenomena.
1.
PARALLELS WITH FOR-PHRASES.
'The analysis of infinitive complements as for-phrases
has been traditional with predicates like wait, hope, anxious,
etc.
1. he was waiting for a good movie
he was hoping for a good movie
he was anxious for a good movie
2. he was waiting to find a good movie
he was hoping to find a good movie
he was anxious to find a good movie
The infinitive phrases and the for-phrases in the examples
above both characterize the object of anticipation for these
1
predicates. Note also that in psuedo-cleft sentences, infinitive phrases can serve as pseudo-clefted objects of for:
-- what we were hoping
We were hoping for Bill to come early,
for was for Bill to come early.
On the basis of such considerations, examples like those
of (2) are generally derived from constructions like those
in (3) :
for [to find a good movie] 1
PP
he was hoping [ for [to find a good movie] 1
3. he was waiting [
PP
for [to find a good movie] ]
PP
Here we are looking to extend this analysis to other infinihe was anxious
[
2
tive phrases.
Consider the following paradigms:
4. John built a robot to entertain his guests
a. what John built was a robot to entertain his guests
b. what John built to entertain his guests was a robot
c. what John did to entertain his guests was build
a robot
5. John built a robot for entertainment
a. what John built was a robot for entertainment
b. what John built for entertainment was a robot
c. what John did for entertainment was buil& a robot
As is evident from the pseudo-cleft examples, which serve to
distinguish three possible readings for (4) and ( 5 ) , there
is considerable similarity between the interpretations
of
-
the infinitive phrase and the interpretations of the forphrase. In (4a,b) and (5a,b), the underlined phrases are
interpreted as designations of the purpose or function of
3
the robot.
In (4a) and (5a), the underlined phrases must
be analysed as complements to a nominal, while in (4b) and
(5b), they are constituents separate from the nominal. In
(4c) and ( 5 c ) , the underlined phrases designate the reason
or rationale for John's action: John's building of the boat
was to entertain his guests, John's building of the boat was
for entertainment.
From the discussion of the last chapter, it is clear
that the infinitive phrase functions as an infinitival
relative clause in (4a), as a purpose clause in (4b), and
for-phrase in (5) can
as a rationale clause in (4c). The serve the same semantic function as each of these clause
types.
Let us now examine the parallelism between objective
for-phrases.
clauses and 6. a. Ann sent John into town to get some groceries
b. what Ann sent John into town for was to get
some groceries
7. a. Ann sent John into town for some groceries
b. what Ann sent John into town for was some
groceries
8. a. John went into town to get some groceries
for was to get some
b. what John went into town groceries
9. a. John went into town for some groceries
-
b. what John went into town for was some groceries
The for-phrases of the free relative clause subjects of the
pseudo-cleft examples (6b) and (7b) are interpreted identically, and the sentences show that either an infinitive phrase
for on such an interor an NP can qualify as the object of pretation. A comparison of (8b) and (9b) indicates the same.
Furthermore, (6) and (8) are ambiguous; the in£initive phrase
can be interpreted either as an objective clause or a rationale clause. This is clearest in ( 6 ) , where either John or
Ann
- can be the understood subject of the infinitive complement. Observe that there is a similar ambiguity in (7): the
for-phrase
-
can be interpreted as designating John's objective
in town, similar to Ann sent John into town after some groceries, in pursuit of some groceries, or it can be interpreted as designating Ann's reason for her action of sending
John into town, similar to Ann sent John into town, intending
to get some groceries in return.
We might add one further case of parallelism of semantic
for-phrases.
function between infinitive phrases and -
After
verbs such as build and design, both infinitive phrases
for-phrases characterize resultant qualities of the
and direct objects.
Consider the following examples:
10. a. we built the car for high durability at high
speeds
b. we built the car to look like an elephant
11. a. we designed the machine for maximal effectiveness at low temperatures
b. we designed the machine to operate noiselessly
With the verb make, such complements are at best questionable, and with verbs like buy, they are impossible.
12. a. ?we made the car for high durability at high
speeds
b. ?we made the car to look like an elephant
*we bought the machine for makimal effectiveness
at low temperatures
b. *we bought the machine to operate noiselessly
In (12) and (13), the a- and b-examples have the same
grammatical status, suggesting that the for-phrase and the
infinitive phrase are alike in their semantic function and
eqnazby compatible or incompatible with the verbs of the
4
sentences.
We have demonstrated, then, that non-clausal forphrases can take on the same semantic functions as each of
the clause types examined in the last chapter (purpose,
rationale, and objective).
forWe will now show that the -
phrases, on their various interpretations, have essentially
the same distributional properties as their clausal counterparts. Specifically, we will argue that the syntactic analysis outlined in the last chapter must be revised to account
for the distributional properties of non-clausal for-phrases
as well as infinitive phrases.
(See pp.31 ff. of Chapter I.)
First of all, we observed earlier that only rationale
clauses could prepose to the beginning of the sentence.
It
is also the case that a for-phrase can prepose only with a
rationale interpretation. Thus, if the infinitive phrase of
(4) is preposed, the sentence has only the reading paraphrased by (4c), assuming non-topicalization intonation
for-phrase of ( 5 ) is
contours. By the same token, if the preposed, the sentence has only the reading paraphrased by
(5c):
14. to entertain his guests, John built a robot
C
15. for entertainment, John built a robot
Second, the various for-phrases manifest the same kinds
of dependency on the matrix predicate as do their corresponding clause types.
Rationale for-phrases are compatible
with volitional actions, conditional predicates (necessary,
sufficient, need), and some of the modals.
(Cf. pp.3sff. of
the last chapter.)
16. John (*accidentally) let the cats out of the room
for some peace and quiet
17. *Bill was short for maximal effectiveness
18. Bill must be short for maximal effectiveness
19. Bill needs to be short for maximal effectiveness
20. it is sufficient for Bill to be short for maximal
effectiveness.
for-phrase is incompatible with the
In (16), the rationale adverb accidentally, which implies lack of volitionality on
the part of the subject. Rationale clauses and for-phrases
connote motivation and, hence, attribute volitionality to the
matrix subject, making sentences like (16) coneradictory if
the adverb accidentally is included. In (17), the subject of
short is not understood as volitional. As a result, (17) is
not acceptable in the same way that a sentence like *Bill
was intentionally short is not acceptable. However, if we
introduce the modal must,-as in (18) or?the conditional
predicates, as in (19) and ( 2 0 ) , the resulting sentences
are grammatical.
for-phrase or
In such sentences, the -
infinitive phrase complement ("rationale"'c1ause)has a
result interpretation:
The for-phrase or infinitive phrase
designates a result which is dependent on a condition defined by the matrix clause in the case of examples like
(18) or by the complements of the conditional predicates
in the case of examples like (19) or (20).
Function-designating for-phrases, like purpose clauses,
show a tighter dependency on the matrix predicate. Thus, such
for-phrases are compatible with many but not all predicates
denoting volitional actions. (See p.35 of Chapter I.)
21. a. Mary built the board for her chess games
b. Mary built the board for Spassky to play on
22. a. Dlaxy boughtLthe baardqfpr her chess games
b,. Mary bought the board for Spassky to play on
23. a. Harold made the stove for his gourmet cooking
b. Harold made the stove for his chef to cook on
24. a. Harold used the stove for his gourmet cooking
b. Harold used the stove to cook on
25. a. ??Nary destroyed the board for the bonfire she
was makins
25.>h. ??Mary destroyed the board for the scouts to burn
repaired the board for her chess games
b. ??Mary repaired the board for Spassky to play on
27. a,* ??Harold painted the stove for his gourmet cooking
b. ??Harold painted the stove for his chef to cook on
28. a. ??Harold cleaned the stove for his gourmet cooking
b. ??Harold cleaned the stove to cook on
(Some of the a-examples in (25-28) may seem to be acceptable
but on a different interpretation of the for-phrase than the
one we are considering in this discussion. For example,
(27a) may be acceptable if interpreted roughly as, Harold
painted the stove in preparation for his gourmet cooking.
In this event, the for-phrase is not interpreted as a designation of the purpose or function of the stove.)
Pur-
pose for-phrases are, in general, compatible with the same
broad classes of predicates as purpose clauses
--
predicates
of transaction, transitive verbs of motion, verbs of creation,
the verb use, etc.
(See p.35
of Chapter I.)
Objective for-phrases like their clausal counterparts,
-
are complements to predicates of motion, such as send, bring,
take,
-go, come, etc.
29. John's parents sent him to Stangord for an
education
John's parents sent him to Stanford to get an
education
30. John went on;to Stanford for an education
John went on to Stanford to get an education
Other verbs, such as hire and train, also take for- phrase
complements:
they hired/trained him to do the job, they
hired/trained him for the job.
The facts concerning preposability and dependency on
the matrix predicate argue that rationale for-phrases, as
opposed to objective for-phrases and purpose for-phrases,
are generated outside the VP (perhaps as daughters of the
S node; see note 5). Another supporting fact is that, ignoring the possibility of preposing, purpose and objective
for-phrases
-
precede rationale for-phrases
--
again, like
their clausal counterparts.
31. a. Harold used his stove for his gourmet cooking
for the thrill of it (Purpose for-PP precedes
rationale for-PP.)
b. *Harold used his stove for the thrill of it for
his gourmet cooking
purpose for-PP.)
(Rationale for-PP precedes
32. a. Bernard took Julia to Lord and Taylor for her
clothes for his own amusement
(Objective
precedes rationale for-PP.)
b. *Bernard took Julia to Lord and Taylor for his
own amusement for her clothes
(Rationale for-PP
for-FP.)
precedes objective -
Yet a fourth consideration is that certain facts about
complement subject control in these clause types, as discussed on pp.37- 40 of the last chapter, seem to pattern after
certain semantic relations that for-phrases enter into with
NP's in the sentence. For example, let's consider again sentence (25) of the last chapter, repeated here "for convenience
as (33).
33. Mort sent his robot to us to get the prize
We observe, as before, that this sentence is ambiguous:
On
one reading, the subject of the infinitive phrase is understood
to be the robot, and the infinitive phrase is an objective
clause. On the other reading, the subject of the infinitive
-
phrase is understood to be Mort, and the infinitive phrase
is a rationale clause. Consider now the following sentence:
34. Mort sent his robot to us for the prize
59
Sentence (34) is likewise ambiguous.
On one reading, the for-
phrase designates the robot's objective, so that there is a
semantic relation established between the robot and for the
prize.
On the other reading, the for-phrase, as a rationale
phrase, is subject-oriented; it designates the motivation for
Mort's action of sending the robot to us.
Thus, the case
where his robot is identified as the controller of the subject
of the infinitive in (33) is matched by the case where his
robot bears a particular semantic relation to the for-phrase
in (34) -- the relation of individual to objective.
Also,
the case where Mort is identified as the controller of the
complement subject in (33) is matched by the case where a
particular semantic relation obtains between the subject Mort
-
and the for-phrase in (34)
--
the relation of an agent to a
motivation.
Turning to purpose clauses, recall that when a purpose
clause appears without an explicit (lexically specified) subject as a complement of the verb buy, the complement subject
is controlled by the matrix indirect object or the matrix
subject, if there is no indirect object.
(We are ignoring
the case where the matrix direct object controls the complement subject, limiting ourselves to the case where the matrix
direct object controls a complement object node.)
the following examples :
Consider
35. Mary bought her daughteri a beautiful doll
to make the others jealous with
-j
to make the
bought a beautiful doll
others jealous with
i
j-
-j
In contrast, the subject of a rationale clause is controlled
by the matrix subject even when an indirect object is present:
37. Maryi bought her daughter a beautiful doll
i to
make the others jealous
38. Maryi bought a beautiful doll
i
to make the
others jealous
Consider now the following example with a for-phrase:
39. Mary bought her daughter a set of rosary beads for
penance
40. Mary bought a set of rosary beads for penance.
(39) and (40) are ambiguous:
the phrase for penance des-
cribes either the purpose of the set of rosary beads or the
reason for the act of buying them (the purpose of the action).
~ h u s ,for penance can be interpreted either as a purpose phrase
or a rationale phrase.
(It can also be interpreted as a
component of the phrase a set of rosary beads for penance,
.
but we will not
:
consider this possibility.)
Now observe the
When
~
(39) is interpreted as a purpose phrase
parallels with the controh phenomena discussed above:
the for-phrase in
(designating the function nf the direct object, a set of
rosary beads), penance is associated with the indirect object
her daughter:
i.e., it i
the daughter's penance that is
being referred to; and in (40), on a purpose phrase interpretation, it is Mary's
that is being talked about.
h
On the other hand, when t e for-phrase is interpreted as a
-
rationale phrase, penance is associated with Mary, the matrix
subject, in both (39) and (40).
(39) and (40) can, on the
The examples discussed h e ~ estrongly indicate that the
control facts for sentences like (35-38) are intimately
connected with some facts about semantic relationships which
for-phrases enter
-
into wikh various NP's, as described in
the discussion of (39) ani (40).
explore certain aspects
04
In the next section we will
these relations between NP's and
for-phrases. For now, we are still simply pointing out the paralJ.&ls
. between
phrases and for-phrases to jus,-
tify the analysis
types discussed in the last
-
chapter as in£initival fod-phrases.
I
A final analogy that we will draw with previous obser-
-
vations about the various clause types is that a for-phrase
62
forming part of a VP focus constituent in a pseudo-cleft
sentence cannot be interpreted as a rationale phrase.
Chapter I, p.40
(See
.)
41. Sam read The Master and Margarita for his own
amusement
a*' what Sam did for his own amusement was read The
Master and Margarita
b. *what Sam did was read The Master and Margarita
for his own amusement
42. Alice played hookey for revenge
~
was play'hookey
a. what Alice did f o revenge
b. *what Alice did was play hookey for revenge
In (41) and (42), the b-sentences
sentences.
cannot paraphrase the a-
Consider further that if we take an example where
for-phrase can be interpreted either as a rationale phrase
the -
-
or an objective phrase, the inclusion of the for-phrase in
the focus constituent serves to isolate the objective-phrase
reading:
43. Marc brought his boss home for a turkey dinner
a. what Marc did for a turkey dinner was bring his
boss home
63
43. b. what Marc did was bring his boss home for a
turkey dinner
(43a) and (43b) cannot be interpreted as paraphrases.
for-phrase
-
The
in (43b) has only the objective sense, and we
understand that Marc's boss is a dinner guest.
In (43a) we
understand that Marc's compensation for bringing his boss
home is a turkey dinner, an interpretation quite different
from (43b). Recall the parallel for objective and rationale
clauses:
only the objective clause can form part of the
focus VP constituent.
Thus, if we take an ambiguous example
such as, Marc broughtlhis .boss*home.to:observe
his--wifeCs
-
behavior, where either Marc (rationale-clause reading) or his
boss (objective-clause reading) can be understood as the complement subject, we find that the relevant pseudo-cleft
example is unambiguous:
44. what Marc did was bring his boss home to observe
his wife's behavior
(44) has only the reading where his boss is the understood
complement subject: i.e., it has only the objective-clause
reading.
Like objective for-phrases, purpose for-phrases can
form part of the VP focus constituent.
45. John made a recorder for his music lessons
a. what John made for his music lessons was a
recorder
b. what John did was make a recorder for his
music lessons
The for-phrase in (45b) has the same interpretation as the
far-phrase in (45a).
-
(Again, there is the further possi-
bility of interpreting a recorder for his music lessons as
a phrase.
In this case, the for-phrase designates purpose
or function, but it is a nominal complement, rather than a
purpose for-phrase.)
To sum up, we have observed in some detail the number
of significant parallels in syntactic and semantic properties
-
between non-clausal for-phrases and the clause types isolated
in the discussion of Chapter I.
Accordingly, I propose the
following revisions in the PS rules given as (32) and (33)
in the last chapter.
6
46.
VP-.3
AUX
Rationale phrases (clausal and nnn-clausal) are introduced by Rule ( 4 6 ) ; objective and purpose phrases are in-
.
troduced by Rule (47)
Rule (48) provides the underlying
structure for rationale, purpose, and objective clauses
(infinitival for-phrases).
In tree diagram form, we get
the following:
AUX
- rationale forXause
,-
1S
A
objective, purpose forphrase
- objective, purpose clau
Thus, on this analysis, the syntax of English distinguishes two positions for for-phrases in deep structure, pro-
-
I
vided by Rules (46) and (47).
*
Furthermore, the for-phrases
can either be clausal or non-clausal.
will be discussed in the next chapter.)
for-phrase
-
(Gerundive for-phrases
The higher-generated
is variously interpreted as a designation of cause,
7
motivation, or result, etc., depending on a variety of factors.
66
Lower-generated for-phrases also show a great deal of latitude in their interpretation (purpose, objective, resultant
quality, etc. )
.
In the next section, we will consider further certain
distinctions between purpose phrases and objective phrases
which will provide the basis for an analysis of control
phenomena in pupose clauses and objective clauses to be
considered in the last parts of the chapker.
2.
On semantic relations between NP's and for-phrases.
In this section, I will consider certain facts related
to the interpretation of lower-generated for-phrases. Specifically, we will give an intuitive description of various
ways in which certain NP's can be semantically related to
for-phrases.
-
Consider the following examples.
1. we bsought the horse to the stables for some hay
2. Bill stopped by Henry's for a rubdown
3. Bill brought John to Henry's for a rubdown
The underlined NP's are semantically related in some way with
the for-phrases of their respective sentences.
In (I),
this relation allows for the inference that our intention
is for the horse to get some hay.
The association between
Bill and the for-phrase of (2) seems to be the same as the
association between John and the for-phrase of (3): In
67
(2),
we infer that Bill is to get the rubdown; in (3), we
infer that John is to get the rubdown.
Of course, the two
cases are distinct in that we infer from (2) that Bill has
volition with respect to his getting a rubdown, whereas, in
( 3 ) , there is no implication of volition on John's part.
In fact, in (3), volition is attributed to Bill with respect
to John's; getting a rubdown.
Thus, Bill is the volitional
individual in both (2) and (3), but, aside from the matter
of volition, Bill bears a certain semantic relation to the
for-phrase in (2) that is equivalent to the semantic relation that John bears to the for-phrase in (3). Moreover,
this semantic relation is intuitively the same as the semanfor-phrase in (1).
tic Zelation between the horse and the Observe that, as with John in (3), there is no implication
of volition on the part of the horse with respect to its
getting some hay.
In fact, as with (2) and (3) volition
is attributed only to the subject of the sentence.
There is reason to believe that the semantic relations
noted above are tied up with the system of thematic relations developed by Gruber (1965) and elaborated by
Jackendoff (1972).
'
(See their works for relevant defini-
tions and discussion.
I will assume familiarity with their
system of thematic relations in what follows.)
Notice
that the underlined NP's in (1-3) all bear the relation of
Theme to the verbs of their respective sentences.
Also,
68
the subjects of the sentences bear the relation of Agent.
We can therefore identify the semantic relation between the
underlined NP'S and the for-phrases as semantic relations
between the Themes of the sentences and the for-phrases.
Also, we can s9y that the Agency of the subjects of the sentences accounts for their volitionality with respect to the
intentions depicted by the -for-phrases.
,
In the case of ( 2 j : ,
Bill is both Theme and Agent of the sentence and accordingly both bears the relevant semantic relation to the forphrase and is volitional with respect to the intention that
the for-phrase depicts.
Observe, incidentally, that the sentences (1-3) are
ambiguous;
-
the for-phrase is also interpretable as what we
have termed a rationale phrase.
The rationale phrase read-
for-phrases are preposed.
ings are brought out if the 4.
for some hay, we brought the horse to the stables
5. for a rubdown, Bill stopped at Henry's
6. for a rubdown, Bill brought John to Henry's
We understand.that the Agent subjects of the sentences are
compensated for their actions by being given some hay in the
case of (4), and by being given a rubdown in the case of
(5) and (6).
Observe that there are no Theme-for-phrase
-
relations here of the type discussed above; the horse is
not understood to be getting the hay in (4) and John is
not understood to be getting a rubdown in (6).
(The
connection between Bill and the for-phrase in ( 5 ) does not
involve the fact that Bill is Theme.)
This is precisely what
we would expect given our syntactic analysis of rationale
phrases.
Since they are generated outside the VP and bear,
consequently, no grammatical relation to the verb, we would
expect them not to be involved in the network of semantic
relations tied to the system of thematic relations.
Of - . ,
course, as we have already seen, the presence of such
phrases requires the attribution of Agency to the subjects
of the sentences in which they occur, but this is the sort
of Agency that is attributed to derived subjects, as can be
seen from the acceptability of a sentence like, for some hay,
my horse would be whipped by anyone.8
The interpretation of the relation between the Theme
for-phrase is subject to variation.
of the sentence and the Contrast, for example, the following cases:
7. John brought the maid to the restaurant for something to eat
8. John sent the maid to the restaurant for some-
thing to eat
In (8), in contrast to ( 7 ) , it is possible to interpret the
for-phrase as
-
a designation of the maid's objective; i.e.,
we can infkr responsibility and volitionality on the part of
the maid with respect to John!s
intention.
(The relation
70
here is similar to the relation of responsibility to be
discussed in Appendix A of this chapter.)
The contrast
between (7) and (8) is brought out in the examples below:
9. a.
John brought the maid to the drugstore for
something to heal herself with
b. *John brought the maid to the drugstore for
something to heal himself with
10. a.
John sent the maid to the drugstore for something to heal herself with
b.
John sent the maid to the drugstore for something to heal himself with
In (lob), the maid is understood to be some sort of agentive intermediary in John's acquisition of something to
heal himself with.
Clearly, the various semantic ielations will have to
be given a more systematic and illuminating analysis.
I
will limit myself here to pointing out an important distinction that any theory of such semantic relations will
have to provide. 9
Consider the differences between the natural interpretations of the sentences below.
11. John brought Mary home for dinner
12. John brought the pizza home for dinner.:
----
Mary and the pizza
I
? I8 L
,
I
I
I
I ,
,1I
Ll,,
(8,
"I,?,
.
,
,
I
are the Themes of their respective sen-
tences, but their semantic relations to the phrase for dinner
are distinct.
Notice that the semantic relation between
the pizza and the for-phrase in (12) is the same as in the
copuJar sentence, $he pizza,is --for-<idinner.On the other -
,
hand, the sentence, Mary is for dinner certainly does not
manifest the same Theme-for-phrase
that (11) does, on its
natural interpretation.10
We will say that the Theme of
for-phrase.
(12) has a functional relation to the -
The for-
phrase can be said to designate the function of the Theme,
in the sense that the Theme is understood to be an object
intended to serve a certain purpose as designated by the
for-phrase.
viously
((12) thus has what we have identified pre-
as a purpose phrase interpretation.
Here, in-
stead of talking about the interpretation of the forphrase, we are talking in terms of the interpretation of
the relation between the Theme and the for-phrase.)
The
Themes of the following examples all have a functional relation to their for-phrases.
13. a. John bought a new car for his trip out west
b. the new car is for John's trip out west
14. a. Tom kept the box for his sewing material
b. the box ig for Tom's sewing material
15. a. Seymour used the knife for slicing salamis
b. the knife is for slicing salamis11
The grammar of English will somehow have to account for
the distinctions in semantic relations between NP's and forphrases that have been noted in this section, since they are
important in the characterization of ambiguities and of differences between the interpretations of sentences.
3.
On control in infinitive for-phrases.
We can see from the last section that lower for-
phrases can be Theme-oriented in the sense that they define
some intention that the Agent has in mind for the Theme.
The
he me-orientation of the for-phrase is reflected in
the fact that the Theme controls the complement subject
when the for-phrase is infinitival.
last section ( p . 6 6
)
Compare (1-3) of the
with the following:
1. we brought the horse- to'the stables to be groomed
and fed
2, Bill stopped by Henry's to get a rubdown
3. Bill brought John to Henry's to get a rubdown
In each case, it is the Theme of the sentence (the underlined NP) that is the understood subject of the infinitive
complement.
Thus, we understand from (1) that the horse is
to be groomed and fed; from (2) that Bill is to get a rubdown; and from (3) that John is to get a rubdown.
We ana-
lyze the infinitive complements Q £ (1-3) as for-PP's, as
73
below:
4. we brought the horse to the stables [ p p for ISto
be groomed and fed]]
5. Bill stopped by Henry's [pp for
[S
to get a rub-
clown] ]
6. Bill brought John to Henry's [ p p for
to get a
rubdown] 1
The analysis of these infinitve complement as for-phrases is
essential to the analysis of the control problem.
Examples
like (1-3) indicate that the facts of complement subject control reflect the semantic connection between Theme-NP's and
for-phrases, which is observable independently. Thus, the
semantic connection between the Theme and the for-phrases
ensures that the Theme is identified as controller of the
complement subject.
Notice that there is no such semantic
connection in the case of rationale phrases, and the Theme
does not function as controller.
For example, if the infini-
tive phrase in (3) is interpreted as a rationale clause, John
is not its understood subject:
7. to get a rubdown, Bill brought John to Henry's
The subject of the rationale clause is understood as Bill.
We have already noted that there is no
semantic connection
between the Theme and a rationale for-phrase.
74
In the last section, we observed how the interpretation of the connection between the Theme and the forphrase could vary.
I would like now to discuss a parti-
cular way in which such vakiation is relevant to matters
of control.
The most interesting treatment that I am aware of of
the relationship between semantic relations and control in
infinitive complements is in Jackendoff (1972).
ter Five of that work, especially pp. 214ff.)
(See Chap-
Offering an
analysis of the control problem, Jackendoff argues that
when the subject of an infinitive complement to a verb is
obligatorily controlled, the position of the controller
in the matrix clause is defined on thematic relations.
Con-
sider the following examples:
8. John promised to leave after the first act
9. John promised Bill to leave after the first act
10. John got to leave after the first act
11. John got Bill to leave after the first act
Promise and get require obligatory control of their complement subjects.
For promise, ( 8 ) and (9) show that the con-
troller is the matrix subject (in this case, John) whether
or not the indirect object (in this case, Bill) is present.
Get assigns control differently; the controller
ject of get, if present, or the subject of get,
is the obif there is
no object.
Jackendoff demonstrates that although there is
get and
a shift in the position of the controller NP with other verbs, the thematic relation of the controller to the
verb remains constant.
John is the Theme of sentence (lo),
but Bill is the Theme of sentence (11).
Thus, we can say
that for the verb get, the controller of the complement is
the Theme of the matrix sentence.
For promise, the controller of the complement subject
is the Source of the matrix sentence.
Since the position
of the Source is always the subject in the case of promise,
the position of the controller does not change when an indirect object is added to (8) to form (9). Promise contrasts with permit, which allegedly has the same correspondence of thematic relations and grammatical relations as
promise.
However, in the case of permit, the complement
subject controller is the Goal of the matrix clause.
Com-
pare (9) with (12).
12. John permitted Bill to leave after the first act
Bill, rather than John, is the understood complement subject in (12).
Jackendoff's analysis assumes a control rule with two
essential arguments: (1) the uncontrolled subject of the infinitive complement and (2) the NP in the matrix sentence
bearing the relevant thematic relation as specified in the
lexical entry of the matrix verb.
The semantic relations
relevant to control in this theory are thematic relations,
i.e., semantic relations between constituents and the verbs
they subcategorize.
A thematic relation is the semantic
interpretation of a grammatical relation.
for-phrases suggests that NP's enter
Our discussion of into semantic relations with constituents other than the
verbs they subcategorize and that these semantic relations
are relevant to the analysis of control phenomena.
While
these semantic relations are not independent of the network
of thematic relations, they are clearly not thematic relations themselves.l2
(See Appendix A.)
Let us consider once more a distinction in Theme-for=
phrase relations which we have already commented on.
13. Tommy brought the chicken horn for supper
(13) is ambiguous; the relationship between the chicken and
the for-phrase can be understood in two ways.
On one inter-
pretation, it is inferred that the chicken is to get something to eat for supper.
The role that the chicken is to
play in the realization of Tommy's intentions is an active
one.
On the other interpretation, it is inferred that the
chicken is to be used as something to make the meal from.
The role that the chicken is to play in the realization of
Tommy's intentions is here a functional one.
That is, the
77
chicken is taken to be an object which is intended to serve
a certain purpose.
The particular purpose it is to serve
for-phrase.
is characterized by the When the for-phrase is infinitival, the ambiguity of
the relation between the Theme and the for-phrase is matched
by an ambiguity of control.
14. Tommy brought the chicken home to eat
a.Tommy brought the chickeni home -i to eat
b.Tornrny brought the chickeni home to eat As (14a) and (14b) indicate, the chicken can be understood
as either the subject or object of eat.
Each of the read-
ings of (14) reflects one of the possible readings of (13),
with (14a) corresponding to the non-functional reading of
(13) and (14b) corresponding to the functional reading of
(13).
Furthermore, notice that just as the copular sentence,
the chicken was for supper isolates the functional interpretation of the Theme-for-phrase
relations, the copular
sentence, the chicken was to eat isolates the "objectdeletion" interpretation of the infinitive; i.e., the chicken is understood uniquely as the object of eat, when the
infinitive phrase figures as a predicative for-phrase in
a copular sentence.
Note, however, that the functional relation between the
Theme and the for-phrase does not correlate only with object
78
deletion.
We saw in the first chapter that the NP in the
purpose clause controlled by the Theme of the matrix sentence is not limited to any particular syntactic position
such as subject or object of the complement.
Consider, in
this regard, the ambiguity of the following examples:
15. the usheri is there
-
to receive the tickets
16. the undercover agentsi are here
- to screen
out undesirables
17. the accountanti is here
-
to look over the books
The semantic relations between the matrix subjects and the
infinitival for-phrases of the above examples can be either
a functional relation or what we might term an intentional
relation. A clue to the ambiguity of these examples lies in
the ambiguity of the term purpose, which can mean either
intention or function.
We can speak of someone's purpose
*(say, in making such and such a statement), or we can speak
of the purpose of an object, such as a chair.
The infini-
tive phrases in the examples (15-17) designate purposes in
the sense of functions on one kind of reading and in the
sense of intentions on the other kind of reading.
Thus, the
infinitive phrase in (15), for example? designates either
the purpose of the usher or the usher's purpose.
On the for-
mer interpretation, there is a functional relation between
the usher and the infinitive phrase, and the infinitive
79
phrase is what we have been calling a purpose clause. (Observe, incidentally, that the ambiguity of (15) is analogous
to the ambiguity of the question, what is the usher there
for?)
NOW observe that when the sbbjects control non-subject
NP's in the infinitive phrases, the sentences are unambiguous, with only a functional relation possible between the
su5ject and the infinitive phrases.
18. the usheri is there for the people to give their
tickets to
-
19. the undercover agentsi are here for criminals to
confess to 20. the accountanti is here for the students to take
lessons from
-
Let us consider what happens when a functional relation does
not obtain between the Theme and the for-phrase.
Notice that
the ambiguit.~of (21) is parallel to the ambiguity of (15).
21. the usher is there for moral support
(21) implies either that the usher seeks moral support (the
for-phrase designating
-
the usher's puapose in being there)
or that the usher provides moral support for others (the
for-phrase designating the usher's function).l3
YII
If we change
the main verb from be to go, a functional relation between
the usher
22.
and for moral support is not possible.
the,t.ushergoes there for moral support
The for-phrase in (22) unambiguously denotes the usher's
purpose, not the purpose of the usher.
In parallel fashion,
go substitutes for be in (15), the Theme-for-phrase
relaif tion is disambiguated:
23.
the usher goes there to receive the tickets
Furthermore, the substitution of go for be in (18) results
in an ungrammatical sentence.
24. *the usher goes there for the people to give their
tickets to
These data clearly show that operation of the rule which interprets or deletes
an NP in a purpose clause under iden-
tity with the Theme of the matrix clause depends on the appropriate semantic relation obtaining between the Theme and
the infinitival for-phrase.
We can illustrate the point further with the adjective
ready.
25.
the patientilis ready for the doctor to operate
on himi
26.
the patienti is ready for the doctor to operate
I n ( 2 5 ) , r e a d y i s a p s y c h o l o g i c a l p r e d i c a t e , much l i k e
for-phrase
e a g e r , a n x i o u s , w a i t and hope (which a l s o t a k e complements), d e s c r i b i n g t h e p s y c h o l o g i c a l s t a t e of i t s subj e c t with respect t o t h e anticipated event described i n the
for-phrase
-
complement.
S i n c e r e a d y i s a p s y c h o l o g i c a l pred-
i c a t e i n t h i s c a s e , i t s s u b j e c t must be an a n i m a t e NP, g i v e n
t h a t we can t a l k a b o u t p s y c h o l o g i c a l s t a t e s o n l y w i t h r e g a r d
t o animate N P ' s .
Thus, we c a n n o t have s e n t e n c e s l i k e ,
27. * t h e tumor i s r e a d y f o r t h e d o c t o r t o o p e r a t e on it
28. * t h e c a d a v e r i s r e a d y f o r t h e d o c t o r t o o p e r a t e on . it
Of
c o u r s e , t h e f a c t s a r e t h e same f o r t h e o t h e r psycholog-
i c a l p r e d i c a t e s r e f e r r e d t o above.
I n ( 2 6 ) , ready i s not i n t e r p r e t e d a s a psychological
p r e d i c a t e ; r a t h e r , it i s d e s c r i p t i v e of t h e s t a t e of a physi c a l o b j e c t w i t h r e s p e c t t o some p r o c e s s it i s t o undergo,
a s c h a r a c t e r i z e d by t h e for-phrase;
I n t h i s case, t h e ani-
macy of t h e s u b j e c t of r e a d y i s i r r e l e v a n t .
Compare ( 2 7 )
with ( 2 8 ) .
28.
t h e tumor i s r e a d y f o r t h e d o c t o r t o o p e r a t e on
t h e c a d a v e r i s r e a d y f o r t h e d o c t o r t o o p e r a t e on
P s y c h o l o g i c a l p r e d i c a t e s do n o t t a k e complements w i t h d e l e t e d
objects.
82
29. *the patient is anxious for the doctor to operate
on
*the patient is waiting for the doctor to operate
*the patient is eager for the doctor to operate on
Observe now that the ambiguity associated with ready is
independent of the control phenomena connected with the foregoing examples.
(30)
Compare (25) and ( 2 6 ) with the ambiguous
.
30.
the patient is ready for the operation
Ready in this example can have either of the senses of the
predicate illustrated in ( 2 5 ) and (26).
The possibility of
an alternative to the psychological-predicate interpretation
allows for examples like, the cadaver is ready for the operation.
Of course, the other psychological predicates men-
tioned above take non-infin2tival-- for-phrase complements:
anxious for an operation, eager for a storm, wait for a catastrophe, hope for a revelation.
However, they do not man-
ifest the ambiguity that ready does. l4
~rucially,psycho-
logical predicates do not allow for a semantic relation between their subjects and their for-phrase coraplements that
correlates with the possibility of complement object deletion.
The semantic relation associated with the predicate
83
ready on its non-psychological interpretation does not seem
to be the same as the semantic relation between a Theme and
for-phrase.
a purpose -
(However, perhaps a more enlightening
relations,
analysis would identify the two Theme-for-phrase
factoring out differences in the semantic functions of pur-
-
-
pose for-phrases and for-phrase complements to adjectives
like-: ready.)
Nevertheless, like the functional relation
this semantic relation
between Theme and purposetfor-phrase,
clearly correlates with the possibility of deleting the complement object.
(This relation also allows for complement
subject deletion, as in, the cadaver is ready to be operated
on, the soup is ready to be served. Note the ambiguity of,
the patient is ready to be operated on.
Thus, the rule in-
volved here works like the purpose clause rule in that the
complement NP to be deleted or interpreted is not fixed at
a pqrticular syntactic position like subject or object.)
Let us consider the correlation a little further.
No-
tice, for example, that the verb ready has only the nonpsychological sense in (31).
31.
the nurse readied the patient for the operation
the patient was readied for the operation by the
nurse
(31) is unambiguous.
We would predict, then, that deletion
of the complement object should be possible, and we find that
this is indeed the case.
(We find, in fact, that it is
+
obligatory. )
32.
the nurse readied the patient for the doctor to
operate on
*the nurse readied the patient for the doctor to
operate on him
Contrast this with the verb prepare which seems to have
both a psychological and non-psychological sense.
33.
Thus,
the nurse prepared the patient for the operation
-
is ambiguous; we could continue it with a phrase like, by
describing to him the triviality of the procedure, compatible with the psychological interpretation, or with a phrase
like, by washing him and giving him a preliminary sedative,
compatible with the non-psychological interpretation.
(For
some reason, the passive version of (33) seems natural only
on the non-psychological interpretation.
tion for this intuition.)
I have no explana-
Predictably, both object-deletion
and full sentence complements are compatible with the verb
prepare.
34.
the nurse prepared the patient for the doctor to
operate on
the nurse prepared the patient for the doctor to
operate on him
85
he psychological readying is associated with the second sen-
tence of (34).
(Note incidentally that the passive version
of the second sentence seems unacceptable: *the patient was
prepared (by the nurse) for the doctor to operate on him.
This is in line with the observation made above that the
passive is compatible with prepare only on its non-psychological interpretation.)
Notice now that prepared, as an adjective, has only the
psychological-predicate sense.
Consider (35), with prepared
as an adjective (as distinct from the interpretation of (35)
as a passive sentence).
35.
the patient was prepared for the operation
Given (35) with this particular interpretation, we cannot
substitute a non-animate NP for the sbbject: *the tumor
was prepared for the operation.
With the adjective prepared,
as with the other psychological predicates, an object-deletion complement is ruled out.
36.
the patient was prepared for the doctor to operate
on him
*the patient was prepared for the doctor to operate
on
Finally, the nominalization readiness has only a psychological reading.
86
37.
the readiness of the patient for the operation was
astounding
*the readiness of the cadaver for the operation was
astounding
the patient's readiness for the operation was astounding
*the cadaver's readiness for the operation was
astounding
As we would predict, the noun readiness cannot take objectdeletion complements.
38.
the readiness of the patient for the doctor to operate on him was astounding
*th& readiness of the patient for the doctor to operate on was astounding
the patient's readiness for the doctor to operate
on him was astounding
*the patient's readiness for the doctor to operate
on was astounding
For a final illustration, let us consider some facts-:
about the verb send.
In ( 3 9 ) , the to-phrase can be inter&-
preted either as a dative phrase or a directional phrase.
39.
John sent Fido to the librarian
-
On the dative-phrase reading, (39) is equivalent to, John
sent the librarian Fido.
Now, if an objective phrase is ad-
to-phrase is unambiguously interpreted as
ded to (39), the a directional phrase.
40. a.
,
John sent Fido to the librarian for his books
b. *John sent the librarian Fido for his books
(40a) is unambiguous, the impossibility of a dative-phrase
interpretation being reflected in the ungrammaticality of
(40b).
(40b), however, may be grammatical if for is taken
to mean ?in exdhange for".
Consider (41):
41. a.
John sent his son to the librarian for an
assistant
b.
John sent the librarian his son for an
assistant
-
The for-phrase of (41a) is interpretable either as a purpose
phrase or as an objective phrase.
On the purpose phrase
reading, there is a functional relation between the Theme,
for-phrase.
his son, and the -
On the objective-phrase read-
ing, there seems to be in this case an intentional relation
for-phrase; the son is to seek out
between the Theme and the an assistant.
The objective-phrase interpretation of the
-phrase correlates with the directional-phrase interpretato-phrase.
tion of the -
Thus, if the to-phrase is unambigu-
ously directional, as in, John sent his son to Boston for an
for-phrase is unambiguously an objective
assistant, the phrase.
.
-
+
With a dative-phrase interpretation of the to-
forphrase, only a purpose-phrase interpretation of the phrase is possible.
In (41b), where the librarian can be
for-phrase is understood only as a puronly a dative, the pose phrase.
Observe, incidentally, that the dative phrase
blocks not only an intentional relation bekween the Theme
and the objective phrase; other Theme-objective-forphrase relations are blocked as well: John sent the papers
to Bill for approval, *John sent Bill the papers for approval.
for-phrase is
(The latter sentence is good only if the taken as a rationale phrase, equivalent to, for approval,
John sent Bill the papers. )
Observe the aonsequences for infinitival for-phrases.
Take examples of objective clauses, such as
42. a.
John sent his son to the librarian to get some
help
b.
John sent his son to the librarian to be
properly trained
If we force a dative interpretation, the sentences are
ungrammatical:
43. a. *John sent the librarian his son to get some
help
b. *John sent the librarian his son to be properly trained
(These sentences, particulargy (43a), would be acceptable on
rationale-phrase interpretations.
One might even stretch
things and say that the sentences are acceptable on pug-_
pose-clause interpretations.
improved as follows:
(43a), for example, might be
John sent the librarian his son to
get some help for her, implying that the librarian was in
need of someone to get some help for her.
But clearly, the
sentences of (43) are at best awkward on such interpretations. )
On the other hand, if we force a directional phrase
interpretation, a purpose-clause complement is impossible-*
44. a.
John sent Rover to Bill for his children to play
with
b.
John sent Bill Rover for his children to play
with
c. *John sent Rover to Boston for the children there
to play with
(44a) is acceptable, if to Bill is understood as a dative
phrase; (44c) seems ungrammatical to me, as we would expect,
although the judgment is somewhat delicate.
90
These examples serve as a further illustration of the
dependency of the control rule effecting complement object
deletion or interpretation on the proper semantic relation
obtaining between the antecedent NP and the phrase
con-
taining the NP to be deleted or--interpreted. In (44a),
Rover can control the complement object because the interto-phrase as a dative phrase allows for
pretability of the fora functional relation between Rover and the infinitival phrase.
In (44c), there is no functional relation between
Rover and the infinitival for-phrase, and the control rule
is blocked.
To sum up, we have examined the relevance of particular
semantic relations between Themes of sentences and forphrase complements to the operation of an interesting kind
of control rule which interprets or deletes NP's in certain
types of infinitive complements under identity with the NP's
serving as Themes of the matrix clauses in which the complements are embedded.
The rule operates in purpose clauses
and in infinitive complements to predicates like ready (on
a non-psychological interpretation).
It effects the dele-:
tion or interpretation of either a complement subject or a
complement object.
In the last chapter, we will speculate
on the nature of this rule.
4.
Chapter summary.
In the first section of the chapter, we gave evidence
that the clause types described in Chapter I are infinitival
for-phrases.
-
That is, they are generated in deep structure
as sentential objects of the preposition for.
In the next
section, we pointed out certain kinds of semantic relations
that obtain between Themes of sentences and for-phrase cornplements generated inside the VP.
The discussion of that
section did little more than to point out the existence
of such semantic relations and was by no means intended
as a definitive analysis of them.
In the third section,
the relevance of these semantic relations to the analysis
of control phenomena was noted.
Here, we made use of the
analysis of purpose and objective clauses as infinitival
for-phrases.
-
We pointed out that the "Theme-orientation''of
certain kinds of for-phrase complements somehow accounts for
the fact that the matrix Theme functions as a complement KP
controller with respect to both objective clauses and purpose-clauses. We then discussed how the availability of
forparticular semantic relations between Theme-NP's and phrase complements governs the applicability of a control
rule which deletes or interprets complement NP's generated
in various syntactic positions.
APPENDIX A
A DIFFERENT KIND OF NP-FOR-PHRASE
RELATION
-
Another kind of NP-for-phrase
relation is discernible in examples like the following:
1. I asked Marie for an invitation
2. we screamed to Nixon for an end to the bombing
3. she begged the committee
for a ruling on the
matter
The underlined NP's are understood to have a role of responsibility with respect to the implementation of the
objective characterized by the for-phrase.
Jackendoff (1972: 34ff.) considers some similar examples and proposes a theory of secondary thematic relations
to account for the relevant semantic relations.
He consi-
ders sentences in which the objects of for are conc~ete
objects, as in
4. Bill asked Alice for a pencil
5. Joe begged Pete for a duckling
(These sentences are not taken from Jackendoff.)
Modeling
our analysis of such examples on Jackendoff's, we would break
down the meanings of (4) and (5) into composite transactions:
In ( 4 ) , we would have a transfer of information (a request)
from Bill to Alice.
Thus, ask would be said to mark its
subject (in this case, Bill) as Source and its indirect
object (in this case, Alice) as Goal.
On Jackendoff's
analysis, there is a secondary transaction, with the pencil
being transferred from Alice to Bill.
Accordingly, the
for-phrase is marked as Secondary Theme; Alice is marked
as Secondary Source; and -Bill
is marked as Secondary Goal.
An analogous analysis would be given for (5).
In actual fact, there is nothing in the sentence which
implies a secondary transfer.
Suppose, for example, that
Alice and Pete are magicians who have mastered the art of
pulling objects out of thin air.
They ask people in the
audience what objects they would like to see appear.
(4)
and (5) would certainly be appropriate to such situations,
and there would be no implication of a possessional transfer of the pencil or the duckling.
It is, however, implicit,
that Alice and Pete-.areresponsible for responding to the
requests that the for-phrases embody.
Many examples that Jackendoff's analysis is supposed
to account for fall the same way, suggesting that his analysis reads factual assumptions about the world into semantic
descriptions.
To take another example, Jackendoff intro-
duces his discussion of secondary thematic relations with a
consideration of the following sentence:
94
6. Esau traded his birthright to Jacob for a mess of
pottage
He maintains that there is a primary transfer of Esau's
birthright from Esau (Source) to Jacob (Goal) and a secondary transfer of a mess of pottage from Jacob (Secondary
Source) to Esau (Secondary Goal).
rect.
But this account is incor-
For example, it is not at all necessary to assume
that Jacob gives the mess of pottage to Esau.
To appreciate
this, consider the fo&lowing modification of ( 6 ) .
7. Esau traded his birthright to Jacob for a modicum
of self-respect
Jacob in (7) has no- active.involvement in Esau's acquiring
a sense of self-respect.
In examples like (2) and (3) it does not make sense to
speak of Nixon and the committee as Secondary Sources, or
of we and she as Secondary Goals.
It makes more sense to
speak of the for-phrases as designating objectives of the
.*
subjects of the sentences and to speak of the underlined NP's
as individuals responsible for the implementation of the
objectives.
This is in effect what Jackendoff does in the case of
infinitival for-phrases with his notion of "assignment of
responsibility".
(However, he does not analyze the infini-
tive complements as for-phrase and does not consider the
95
similarity between the infinitive and non-infinikkve cases.
The focus of his discussion is the role of "assignment of
responsibility" in the identification of the complement subject controller.)
He suggests that when the complement subject is optionally controlled, the selection of a controller should be
free within the constraints imposed by pronominalization.
He discusses this hypothesis in relation to the verbs
scream and shout, which point up the existence of further
parameters in the determination of the complement subject
controller.
Consider these examples:
8. *I screamed to go
9.
I screamed to Bill to go
10.
Idscreamed to Bill for Harry to go
11.
I screamed to be allowed to go
12.
I screamed to Bill to be allowed to go
13.
I screamed to Bill for Harry to be allowed to go
(10) and (13), which have lexically specified complement subjects, show that the complement subject is not obligatorily
controlled in the case of scream.
This means that, accord-
I or Bill should be
ing to Jackendoff's hypothesis, either -
able to control the complement subject.
Notice, however,
I cannot control the complement subject in (8) and (9),
that -
while Bil1:cannot control the complement subject in (12).
The examples clearly show that the controller in the matrix
clause is not fixed for the verb scream; yet in (9) and (12),
there is no ambiguity of control.
Jackendoff argues that a verb can mark a particular NP
in the matrix clause as Agent of the infinitive complement.
When the complement already has an Agent subject, the "ma-trix Agent", as marked by the verb, is understood as the
indirect Agent of the complement, i.e., as an individual who
influences the action of the complement temporally or
causally prior to the efforts of the Agent subject of the
complement.
In (lo), e.g., Bill is understood to be-in a
position to do something to br,$ng it about that Harry will
go.
In this way, Bill is identifiable as an indirect object.
If an empty complement subject is Agentive, it is con-
trolled by the NP marked as matrix Agent by the verb, as in
(9).
(To account for the impossibility of (8), it must be
specified that, for verbs like scream and shout, an Agent
complement subject can be controlled only by the NP in the
matrix sentence marked as Agent of the complement by the main
verb.
This is not the case with all verbs:
If pay or beg
is substituted for scream in (8)) the sentence becomes grammatical.)
If the Subject of the complement is non-Agentive,
it cannot be controlled by the matrix Agent, as in (12).
When we speak of the Agency of the complement subject, we
do not mean deep Agency (a thematic relation), but the Agency
I
! ~ l , , . / , , , . , ,i,
which is attributable to derived subjects.
"
l
l
I
,
,
,
, ,I
.I,,,
In a sentence
like (14)
14. 1 screamed to Bill to be examined by a competent
physician
-
control of the complement subject can go to Bill, and perhaps
also to I.
(I personally find it very difficult to get
clearcut judgments about sentences with scream, including
examples like (14) and the paradigm (8)- (13). )
The subject
of be examined can optionally be interpreted Agentively,
as in Bill managed/intended to be examined by a competent
physician.
-
This allows for the case where Bi1l:is complement
subject controller in (14). It may also be possible that
-I
can control the complement subject in this sentence,
tor-
relating with a non-Agentive interpretation of the complement
subject.
If it is in fact the case that I cannot control the
complement subject, then we must stipulate that, in the
presence of the matrix Agent, the complement subject must be
interpreted Agentively if it is possible for it to be so
interpreted.
Note that we must say, "in the presence of the
matrix Agent," because the sentence, I screamed to be examined by a competent physician is grammatical, with I, of
course, the complement subject controller.
Verbs of the scream-type take for-phrase complements.
If we analyze the infinitive complements in the paradigm
98
for-phrases, we could recast the concept of
(8)-(13) as for"matrix agent" as a relation between a matrix NP and a phrase complement, thereby accommodating examples like
5
Thus, in both of the sentences below
15.
we screamed to Nixon for an end to the bombing
16.
we screamed to Nixon for the generals to end
the bombing
there is the implication that Nixon is assumed to be in a
position to do something to bring about an end to the bombing.
Compare further the a- and b- sentences of (17-19):
17.<>a.I asked Marie for an invitation
b. I asked Marie to invite me to the party
18. a. she begged the Committee for a ruling on the
matter
b. she begged the Committee to be allowed to speak
her mind
19. a. they shouted to me for a pail of water
b. they shouted to me
- to get them a pail of water
In these examples, the underlined NP's are assumed to be in a
position to respond to the--requestscharacterized by the
for-phrases.
-
APPENDIX B
REMARKS ON E Q U I I N PURPOSE CLAUSES
Aside from t h e c o n t r o l - r u l e d i s c u s s e d i n t h e main p a r t
of t h e c h a p t e r , t h e r e i s a r u l e of Equi which o p e r a t e s i n
purpose c l a u s e s , when t h e m a t r i x theme c o n t r o l s a complement
object.
T h i s Equi r u l e o p e r a t e s i n example ( 2 ) ,
1. P e t r r e n t e d t h e p l a n e i f o r S a s h a + t o t a k e
-
to
Moscow
2. Petr
j
rented t h e planei
-j
t o take
- t o Moscow
making P e t r t h e u n d e r s t o o d s u b j e c t of t h e complement.
W e o b s e r v e d e a r l i e r t h a t t h e r e w a s a l s o a n Equi r u l e f o r
r a t i o n a l e c l a u s e s and t h a t i t o p e r a t e d d i f f e r e n t l y from t h e
Equi r u l e f o r purpose c l a u s e s :
The c o n t r o l l e r N P i n r a t i o n -
a l e - c l a u s e Equi i s always t h e m a t r i x s u b j e c t , w h i l e t h e cont r o l l i n g NP i n p u r p o s e - c l a u s e Equi i s sometimes t h e m a t r i x
s u b j e c t , sometimes t h e m a t r i x i n d i r e c t o b j e c t .
3. P e t r i rented t h e plane
-i
4 . P e t r i r e n t e d Sasha t h e p l a n e
Compare:
t o t a k e it t o Moscow
- t o t a k e it t o
Moscow
5. P e t r i r e n t e d t h e p l a n e
- t o t a k e a < * t oMOSCOW
6 . P e t r r e n t e d Sashai t h e p l a n e
Moscow
- t o t a k e $,"to
100
The operation of the complement subject control rule for
rationale clauses remains unaffected by the introduction of
the indirect object.
(See (3) and ( 4 ) ; )
On the other hand,
the controller of the subject of the purpose clause shifts
to the indirect object when it is added.
(See ( 5 ) and (6). )
In the purpose clause examples above, the complement
subject is an Agent.
Now, in the matrix sentence, there is
generally some NP which designates the individual responsible
for the carrying out of the intention characterized by the
for-phrase.
-
Thus, in a sentence like, John gave the tent to
Bill for hunting trips, it is understood that Bill makes
use of the tent in hunting trips.
When the Agent-subject of
a purpose clause is not lexically specified, it is understood
as coreferential with the matrix NP designating the responsible individual.
This is reminiscent of the matching prin-
ciple proposed by Higgins in his dissertation. (Higgins
(1973: 182-3)) l5
Since the Source-Goal transference in the possessional
sense is essentially a transference of control over the
Theme, the Goal NP is typically identified as the understood
complement subject.
Thus, consider the following examples:
7. Bill promised Johni the blue fountain pen - to
do his homework with
8. Billi borrowed the blue fountain pen from John
to do his homework with
-
101
B i l l l e n t t h e b l u e f o u n t a i n pen t o Johni i
9.
t o d o h i s homework w i t h
Now, t h e s u b j e c t o f ' t h e p u r p o s e c l a u s e i s u n d e r s t o o d t o be
e x e r c i s i n g c o n t r o l over t h e b l u e f o u n t a i n pen, i n t h a t t h e
s u b j e c t m a n i p u l a t e s t h e pen a s a n i n s t r u m e n t w i t h which h e
d o e s h i s homework.
The-way t h e E q u i o p e r a t i o n works h e r e i s
t o i d e n t i f y t h e NP which e x e r c i s e s c o n t r o l o v e r t h e f o u n t a i n
pen i n t h e p u r p o s e c l a u s e ( t h e complement s u b j e c t ) w i t h t h e
UP i n t h e m a t r i x c l a u s e which i s u n d e r s t o o d , on t h e b a s i s of
t h e t h e m a t i c r e l a t i o n s , t o be r e s p o n s i b l e f o r t h e u s e of
t h e f o u n t a i n pen w i t h r e s p e c t t o t h e i n t e n t i o n c h a r a c t e r i z e d
by t h e p u r p o s e c l a u s e .
When t h e s u b j e c t o f t h e p u r p o s e
c l a u s e i s l e x i c a l l y s p e c i f i e d , t h e m a t r i x NP i s u n d e r s t o o d
a n a n i n d i r e c t Agent:
B i l l borrowed t h e f o u n t a i n pen from John f o r
10.
..
S a l l y t o d o h e r homework w i t h
B i l l i s a n i n d i r e c t Agent i n ( 1 0 ) ; he i s u n d e r s t o o d t o b r i n g
it a b o u t t h a t S a l l y c a n d o h e r homework w i t h t h e f o u n t a i n pen.
Now l e t u s c o n s i d e r some f u r t h e r e x a m p l e s :
11. a .
I s o l d my f i l e c a b i n e t
b . *I s o l d my o l d f i l e c a b i n e t t o k e e p p e r s o n a l p a = .
pers i n
c.
I s o l d B i l l my o l d f i l e c a b i n e t
102
11. d .
I s o l d B i l l my o l d f i l e c a b i n e t t o keep p e r -
sonal papers i n
( l l b ) i s o u t b e c a u s e t h e m a t r i x c l a u s e d o e s n o t p r o v i d e an
N P which i s u n d e r s t o o d t o e x e r c i s e c o n t r o l o v e r t h e Theme
w i t h r e s p e c t t o t h e purpose c l a u s e .
T h i s N P c o u l d o n l y be
-
t h e Goal of s e l l , s i n c e t h e a c t of s e l l i n g i n v o l v e s t h e
r e l i n q u i s h i n g of c o n t r o l and r e s p o n s i b i l i t y o v e r t h e s o l d
object.
The s i t u a t i o n i s remedied i n ( l l d ) ; t h e i n d i r e c t
o b j e c t of s e l l f u n c t i o n s a s Goal and i s i d e n t i f i e d a s t h e
u n d e r s t o o d complement s u b j e c t .
But t h e r e i s a problem h e r e :
I f t h e Equi r u l e i s o p t i o n a l ( i . e . ,
i f t h e purpose c l a u s e
s u b j e c t c a n be l e x i c a l l y s p e c i f i e d ) , we might e x p e c t ( l l b )
t o become grammatical i f we add a s u b j e c t t o t h e purpose
c l a u s e , assuming t h a t ( l l b ) i s o u t f o r l a c k of a s u i t a b l e
c o n t r o l l e r f o r t h e complement s u b j e c t .
This expectation,
however, i s n o t borne o u t :
12.
*I s o l d my f i l e c a b i n e t f o r B i l l t o keep h i s
p e r s o n a l p a p e r s .i n
F u r t h e r more, n o t i c e t h a t we can have a l e x i c a l l y s p e c i f i e d
s u b j e c t i n t h e purpose c l a u s e i n t h e f o l l o w i n g c a s e , where
t h e r e i s a l s o a m a t r i x Goal:
13.
I s o l d B i l l my o l d f i l e c a b i n e t f o r h i s s e c r e -
t a r y t o keep h i s p e r s o n a l p a p e r s i n
103
Thus, it a p p e a r s t h a t t h e p r e s e n c e of a p u r p o s e c l a u s e makes
o b l i g a t o r y t h e p r e s e n c e of a n i n d i r e c t o b j e c t f o r s e l l .
Compare s e l l w i t h t h e v e r b o f f e r , which a l s o h a s a n AgentS o u r c e s u b j e c t and a Goal i n d i r e c t o b j e c t .
1 4 . a . *Mary o f f e r e d h e r o l d h a t
b.
*Mary o f f e r e d h e r o l d h a t t o s e l l
c.
Mary o f f e r e d u s h e r o l d h a t
d.
Mary o f f e r e d u s h e r o l d h a t t o s e l l
(14a) i n d i c a t e s t h a t o f f e r , l i k e g i v e , r e q u i r e s an i n d i r e c t
P
object.
Thus,
( 1 4 b ) i s o u t , i t would a p p e a r , b o t h b e c a u s e
o f f e r h a s no i n d i r e c t o b j e c t and b e c a u s e t h e s u b j e c t of t h e
complement h a s n o t s u i t a b l e c o n t r o l l e r .
The a d d i t i o n o f
the indirect object u s i n (14d) remedies both t h e s e d e f e c t s .
N o t i c e , however, t h a t t h e f o l l o w i n g s e n t e n c e i s g r a m m a t i c a l :
15.
Mary o f f e r e d h e r o l d h a t f o r t h e l a d i e s t o s e l l
T h a t i s , when o f f e r t a k e s a p u r p o s e c l a u s e complement, t h e
r e q u i r e m e n t of a n i n d i r e c t o b j e c t i s suspended; a n i n d i r e c t
object is optional:
16.
Mary o f f e r e d u s h e r o l d h a t f o r t h e l a d i e s t o
sell
Thus, ( 1 4 b ) i s ungrammatical o n l y f o r t h e l a c k of a s u i t a b l e
c o n t r o l l e r f o r t h e complement s u b j e c t .
T h i s c o n t r a s t between
o f f e r and s e l l h o l d s , a s we would e x p e c t , even i f t h e purpose
phrase i s a non-infinitival
17. a .
for-phrase:
Mary o f f e r e d h e r c a r f o r t h e needed t r a n s portation
b.
Mary o f f e r e d them h e r c a r f o r t h e needed t r a n s portation
c . *Mary s o l d h e r c a r f o r t h e needed t r a n s p o r t a t i o n
d.
Mary s o l d them h e r c a r f o r t h e needed t r a n s p o r -
tation
What we f i n d , t h e n , i s t h a t t h e p a t t e r n s of t h e m a t i c
r e l a t i o n s work t o d e f i n e some m a t r i x N P a s an Agent o v e r t h e
purpose c l a u s e .
An u n c o n t r o l l e d Agent complement s u b j e c t
i s matched w i t h t h i s m a t r i x Agent, a s i n examples ( 7 - 9 ) above.
I t a p p e a r s t h a t some v e r b s l i k e s e l l r e q u i r e t h e p r e s e n c e
of a r e s p o n s i b l e Agent o v e r t h e purpose c l a u s e .
( I l b ) w i t h ( l l d ) and ( 1 7 c ) w i t h (17d) . )
(Contrast
This requirement
i s i n d e p e n d e n t of t h e need f o r a s u i t a b l e c o n t r o l l e r f o r t h e
s u b j e c t of t h e p u r p o s e c l a u s e .
I n a d d i t i o n , some v e r b s l i k e
o f f e r r e q u i r e e i t h e r a Goal p h r a s e o r a purpose p h r a s e , and
t h e s e can cooccur.
are in.)
(Thus, ( 1 4 a ) i s o u t , b u t ( 1 4 c ) and ( 1 5 )
Give does n o t behave l i k e o f f e r i n t h a t t h e p r e s -
ence of a purpose p h r a s e does n o t suspend t h e r e q u i r e m e n t of
an i n d i r e c t o b j e c t :
I gave B i l l my c a r f o r Johnny t o d r i v e ,
P I gave my c a r f o r Johnny t o d r i v e .
I t i s conceivable t h a t
105
t h e d i f f e r e n c e between t h e s e two v e r b s i s t h a t t h e r e q u i r e ment o f a n i n d i r e c t o b j e c t i s a m a t t e r of s e m a n t i c s f o r
o f f e r and a m a t t e r of s y n t a x f o r g i v e .
That i s , o f f e r
r e q u i r e s a p h r a s e of a c e r t a i n s e m a n t i c t y p e , w h i l e g i v e
i s s u b c a t e g o r i z e d f o r t h e o b l i g a t o r y p r e s e n c e of an i n d i r e c t
o b j e c t NP.
I n any e v e n t , t h e E q u i r u l e f o r p u r p o s e c l a u s e s works
a s we have d e s c r i b e d
--
matching t h e u n c o n t r o l l e d Agent
s u b j e c t of t h e p u r p o s e c l a u s e w i t h t h e m a t r i x Agent of t h e
p u r p o s e c l a u s e d e f i n e d on t h e t h e m a t i c p a t t e r n s of t h e
m a t r i x s e n t e n c e . 16
FOOTNOTES
CHAPTER I1
1.
Notice that in the examples with 'wait
the for-phrase
and infinitive phrase are ambiguous in identical fashion;
they each can characterize either the object of the wait
or the reason or rationale for the act of waiting.
Thus,
consider the parallel ambiguities of we waited patiently
for a large sum of money and we waited patiently to make
up for what we had done.
The rationale reading is brought
-
out if the for-phrase and infinitive phrase are preposed:
for a large sum of money, we waited patiently; to make up
for what we had done, we waited patiently.
We will be
considering a number of such parallels in the interpreta-
-
tion and syntactic behavior of infinitive phrases and forphrases throughout this section.
2.
Recall that we noted in passing in the last chapter that
when purpose clauses are pseudo-clefted, the preposition
for shows up
-
in the free relative subject of the pseudo-
cleft sentence:
i.
-
what Carol bought a rack for was to hang coats
This is clearly evidence for the analysis we are advancing
in this section.
In addition, consider the following facts.
It has been observed, e.g., by Emonds, that sentential complements cannot be clefted.
ii. a.
Consider the following examples:
John intended for Mary to leave late
b. *it was for Mary to leave late that John
intended
iii. a.
Max promised to be a good boy
b. *it was to be a good boy that Max promised
iv. a.
Julia said that the cops were harrassing her
b. *it was that the cops were harrassing her that
Julia said
However, at least for a significant number of speakers of
English, purpose clauses and rationale clauses are acceptable in focus position in cleft sentences.
v. a.
b.
Sally bought the piano for Todd to practice on
it was for Todd to practice on that Sally
bought the piano
vi. a.
Sally bought the piano to persuade Todd of her
affections
b.
it was to persuade Todd of her afEections that
Sally bought the piano
If these observations are accepted as correct, a plausible
explanation for the difference between (ii-iv) and (v-vi)
might be that the infinitive complements in (v-vi) are underlyingly prepositional phrases and not simply sentential complements.
This explanation works whether we account for the
facts in (ii-iv) by appealing to ~monds'sanalysis of
cleft sentences (Emonds (1970:113-114; 163-172)) or to
Ross's condition blocking structures with internal S nodes
exhaustively dominated by NP's.
Objective clauses, incidentally, do not seem acceptable in cleft focus position.
vii. a.
John sent Ann to buy herself some new shoes
b. *it was to buy herself some new shoes that
John sent Ann
viii. a.
Ann sent John home to take care of the baby
h. ?it was to take care of the baby that Ann
sent John home
But it seems that in general objective for-phrases (which
will be discussed below) are questionable in cleft focus
position.
ix. a.
John sent Ann for some new shoes
b. *it was for some new shoes that John sent Ann
x. a.
Ann sent John home for the baby's bottle
b. ?it was for the baby's bottle
John home
that Ann sent
To pursue the matter of clefting just a bit further,
consider the fact that (vb) and (vib) are ungrammatical if
for is
-
added at the ends of the sentences:
xi.
*it was for Todd to practice on that Sally bought
the piano for
xii.
*it was to persuade Todd of her affection that
for
Sally bought the piano -
for or the entire forfor-phrases, either the object of With phrase can be clefted:
xiii. a. Sally bought the piano for Todd's practice
sessions
b. it was Todd's practice session'that Sally bought
the piano for
c. it was for Todd's practice sessions that Sally
bought the piano
xiv. a. Sally bought Todd the piano for her own amusement
b.?it was her own amusement that Sally bought Todd
the piano for
c. it was for her own amusement that Sally bought
Todd the piano
If purpose clauses and rationale clauses are introduced by
the rule PP -3 P
S f the ungramaticality of (xi-xii) is
110
accounted for by the prohibition a g a i n s t clefting sentential
complements, however this is to be stated.
(If Ross's
formulation is accepted, the phrase structure rule must
be changed so that an NP dominates the sentential complement: (a) PP+
P
NP, (b) NP+
S. )
In (xi-xii), the senten-
tial objects of for, rather than the entire for-phrases,
have been clefted.
3.
When the for-phrases and infinitive phrases designate
the purpose or function of some object, they can be predicated of that object:
i.
ii.
the robot was for entertainment
the robot was to entertain guests
(i), notice, is a valid inference from (5 arb) and (ii) a
valid inference from (4 a, b)
4.
.
Notice that the ungrammaticality of (13a) can be reme-.
died if we add the possessive pronoun its before maximal
effectiveness:
i.
we bought the machine for its maximal effectiveness at low temperatures
But in such a case, the for-phrase is interpreted quite dif-
ferently: in (i), rather than designating a resultant
111
for-phrase describes a quality-sf the machine
quality, the that motivates its purchase.
If we alter (10a) and (lla)
in the same way, the sentences become ungrammatical:
ii. *we built the car for its durability at high
speeds
iii. *we designed the machine for its maximal effectiveness at low temperatures.
Such examples will come up again at a later point.
5.
Observe that these phrases (both the infinitive phrase
for-phrase), on a rationale interpretation, can be
and the interposed between the subject and the verb:
i.
ii.
John, to entertain his guests, built a robot
John, for entertainment, built a robot
Tanya Reinhart (personal communication) has informed me that
according to Jackendoff (1972), such facts indicate that
the phrases are generated as daughters of the S node.
Jackendoff (1972, Chapter Three).)
(See
In the syntactic analysis
of the last chapter, we generated rationale clauses as
daughters of the node
(=PredP).
If Jackendoff's analysis
is correct, our syntactic analysis will have to be accordingly revised.
However, absolutely nothing in my arguments
hangs on-this paint. In the last chapter, as here, I have
wanted only to establish that rationale clauses and forphrases are generated outside the VP, at some higher level
--
be it Pred P or S.
6.
See Note 5.
7.
The causal interpretation of the for-phrase is evident
in examples like, for no-apparent reason, the plant withered
away.
Motivation, of course, seems intimately connected
with causality, adding the element of will on the part of
an agent.
The result interpretation of the for-phrase oc-
curs, as we have seen before, with modals and the conditional predicates.
Observe, incidentally, that the phrase in order (for
x)
to can introduce infinitives of result only on the root
sense of the modal.
Thus, consider the root-epistemic
ambiguity of the following.
i.
Ben must be quite handsome to attract a girl like
Mary
The epistemic interpretation of the modal can be brought out
if we add a phrase like, as much as he has to the sentence:
ii.
Ben must be quite handsome to attract a girl like
Mary as much as he has
113
If we add in order to, we disambiguate (i) and make (ii)
unacceptable:
iii.
Ben must be quite handsome in order-to attract
a girl like Mary
iv. *Ben must be quite handsome in order to attract a
girl like Mary as much as he has
Must, in (iii), can only have the root sense.
8.
In this sentence, would (in the sense of "be willing1')
also marks the subject as Agent.
9.
Sentences of the following type seem interesting:
i. the judges chose the chihuhua for first prize
This sentence is clearly ambiguous.
On one reading, it is
understood that the chihuhua is being given as a prize, as
in the chihuahua is first prize, or even, the chihuahua is
for first prize.
On the other reading, it is understood that
the chihuahua is to be awarded first prize.
Somehow, the
grammar must provide for this ambiguous interpretation of
the semantic relation between the chihuahua and the forphrase.
Notice that with other verbs, the relation is not
ambiguous.
Consider, e.g., (ii) and (iii).
ii. the judges used the chihuahua for first prize
iii. the judges bought the chihuahua for first prize
In both of these examples, it is understood that the chihuahua is being given as first prize, and the alternative interpretation of the relation available for (i) is not possible
here.
Furthermore, a proper theory of semantic relations
i
will have to allow for further nuances such as appear in
iv.
the judges took the chihuahua for first prize
(iv) is ambiguous; take can be interpreted literally or
figuratively.
On the literal interpretation (cf. the judges
grabbed the chihuahua for first prize), the relation between
for-phrase is similar to the relation
the chihuahua and the in (ii) and (iii) in that it is understood that the chihuahua
is intended to be a pkize, but intuitively there are differences.
On the figurative interpretation, (iv) can be
paraphrased roughly as, the judges took the chihuahua t o ,
be first prize.
(Cf., the judges mistook the chihuahua
for first prize.)
10.
It is important to say, "on its natural interpretation."
Notice that there is an unlikely, but perfectly possible,
interpretation for (11) on which the semantic relation between Mary and for dinner is the same as in the sentence,
Mary is for dinner.
On this reading, we would infer that
the diners are cannibals.
On this interpretation of (ll),
-
the Theme-for-phrase relation is the same as in (12); in
both cases the relation is functional.
(See text above.)
The ambiguity that (11) has is perhaps shown more plausibly with the following example:
i. John brought some chickens home for something
to eat
It can be understood either that the chickens are to do the
eating (cf. the first interpretation of (11)) or that the
chickens are to be eaten (cf. the functional interpretation
of (11)).
11.
See Faraci (1973) for a reconsideration of use-con-
stxuctions.
I argue there that the infinitive phrases of
many of the famous sentences of Lakoff (1968) are to be
-
analyzed as purpose for-phrases.
Obviously, for-phrases which serve as predicates
in copular sentences are not always interpreted as functiondesignating.
Benefactive for-phrases, for example, can
also be predicative:
is for Alex.
I built the house for Alex, the house
Perhaps it will turn out that benefactive and
for-phrases share specific features of their interpurpose pretation, that allow them both to be predicative.
116
12.
Actually, the possibility is implicit in some of Jacken-
doff's discussions that semantic relations between certain
matrix N P 1 s and the infinitive complement are relevant
He suggests at various points that
ko3the control problem.
infinitive complements bear thematic relations to the verbs
that take them, but he does not explore the possible ramifications of this idea.
Conkider the following passage from his book
(Jackendoff (1972:215))
.
Toward [a theory of the difference between promise
get with respect to complement subject control],
and notice the similarity in the following sets of
examples.
(5.136) Joe got to Philadelphia
Frank got Joe to Philadelphia
(5.137) Joe got furious at Henry
Frank got Joe furious at Henry
(5.138) Joe got to wash the dishes
Frank got Joe to wash the dishes
(to7
(5.139)?Joe kept
Frank kept Joe in his room
(5.140) Joe kept at the job
Frank kept Joe at the job
(5.141) Joe kept working on the problem
Frank kept Joe working on the problem
The interesting thing about these examples is that
the switching of understood complement subjects in
(5.138) and (5.141) is exactly parallel to the
switching of attribution of the adjectives and
locatives in the rest of the examples.
And this
swit~hing~in
turn is exactly parallel to the switching of attribution of motion in (5.142).
(5.142) The ~ o c krolled away
Bill rolled the rock away
In other words, we appear to be dealing with a manifestation of the system of thematic relations introduced in Chapter 2.
In the last sentence, Jackendoff is clearly handwaving.
account raises a number of puzzling questions.
His
For example,
he goes on to assert that the PP of (5.136) and the AP of
get.
(5.137) bear the thematic relation ,of Goal to -
In
(5.139) and (5.140) , the PP's are analyzed as Locatives.
What bearing does this have on the attribution of adjectives
and locatives in these sentences? Are we to infer that the
infinitive complement in (5.138) is a Goal and that the
participial phrase in (5.141) is a Locative?
If so, does
118
t h i s have any b e a r i n g on t h e d e t e r m i n a t i o n o f t h e u n d e r s t o o d complement s u b j e c t ?
C e r t a i n l y , it would n o t seem
p o s s i b l e t o h a v e a v e r b which i s l i k e get i n that the infinit i v e complement f u n c t i o n s
as
a Goal p h r a s e b u t u n l i k e get
i n t h a t t h e s u b j e c t of t h e complement i s n o t c o n t r o l l e d by
t h e Theme o f t h e m a t r i x c l a u s e .
It
is
conceivable, then,
t h a t i n c e r t a i n c a s e s t h e p o s i t i o n o f t h e complement subj e c t c o n t r o l l e r i s d e f i n e d on t h e m a t i c p a t t e r n s , i . e . ,
on
s e m a n t i c r e l a t i o n s which a r e t h e m s e l v e s d e f i n e d on t h e m a t i c
relations.
The i n t e r p l a y o f t h e m a t i c r e l a t i o n s , t h e i r
p a t t e r n s , and r u l e s of a t t r i b u t i o n and c o n t r o l i s n e v e r made
c l e a r i n Jackendoff's discussion.
S p e c i f i c a l l y , he does
n o t d e a l w i t h t h e p o s s i b l e r e l e v a n c e t o t h e c o n t r o l problem
o f a s s i g n i n g t h e m a t i c r e l a t i o n s t o t h e i n f i n i t i v e complements.
13.
I w i l l mention h e r e i n passing a f u r t h e r ambiguity.
The s e h t e n c e , t h e g u a r d i s h e r e f o r y o u r p r o t e c t i o n , c a n
imply e i t h e r t h a t t h e g u a r d i s a p r o t e c t o r ( f u n c t i o n a l r e l a for-phrase) o r t h a t t h e g u a r d ' s being
t i o n between Theme and h e r e e n s u r e s t h a t one i s p r o t e c t e d , analagous t o , t h e l i o n
i s i n h i s cage f o r your p r o t e c t i o n .
The a m b i g u i t y i s s t r u c -
t u r a l , a s c a n b e s e e n from t h e f a c t t h a t t h e for-phrase i s
p r e p o s a b l e on t h e s e c o n d r e a d i n g :
f o r your p r o t e c t i o n , t h e
guard is h e r e ; f o r your p r o t e c t i o n , t h e l i o n i s i n h i s cage.
14.
T h e r e L i s a t h i r d possible i n t e r p r e t a t i o n f o r ready,
d e f i n i n g some i n h e r e n t s t a t e of i t s s u b j e c t .
(In the
o t h e r n o n - p s y c h o l o g i c a l i n t e r p r e t a t i o n of r e a d y , t h e r e i s
a n i m p l i c a t i o n of some kind o f , p r e p a r a t i o n of t h e s u b j e c t
for-phrase.)
of r e a d y w i t h r e s p e c t t o what i s d e f i n e d by t h e I n t h e s e k i n d s of examples, o n l y complement s u b j e c t cont r o l is possible.
R e l e v a n t examples a r e , t h e f l o w e r s a r e
r e a d y t o bloom, h e r t e e t h a r e r e a d y f o r a c l e a n i n g , t h e tumor
i s r e a d y t o be o p e r a t e d on, t h e soup i s r e a d y t o be s e r v e d .
Note t h a t t h e l a s t two examples a r e ambiguous; a s i d e from
t h e i r "inherent s t a t e " readings, they a r e a l s o i n t e r p r e t a b l e a s p a s s i v e v e r s i o n s o f , t h e tumor i s r e a d y t o o p e r a t e
on and
-
t h e soup i s r e a d y t o s e r v e , r e s p e c t i v e l y .
There a r e
f u r t h e r examples l i k e , John i s r e a d y t o q u i t h i s job, which,
a s i d e from t h e p s y c h o l o g i c a l - p r e d i c a t e
interpretation dis-
c u s s e d i n t h e t e x t , have a n o t h e r i n t e r p r e t a t i o n which i s
b r o u g h t o u t i n c o n k e x t s l i k e , John g o t s o a n g r y a t h i s b o s s
t h a t he was r e a d y t o q u i t h i s job.
Here, r e a d y means
a.
r o u g h l y , "on t h e v e r g e of d o i n g . . . " .
15,
I t h a s been o b s e r v e d I e . g . ,
Jackendoff
(1972) and Higgins
( 1 9 7 3 ) ) t h a t t h e u n d e r s t o o d s u b j e c t of t h e complement of a
n o m i n a l i z a t i o n must be d e t e r m i n e d on t h e b a s i s of t h e themat i c p a t t e r n s i n t h e sentence.
Thus, c o n s i d e r t h e s e n t e n c e s
below ( J a c k e n d o f f (1972: 2183; H i g g i n s (1973: 1 8 2 ) ) .
i. Mary gave Alex permission to go
ii. l~lary gave Alex a promise to go
Here, the underlined nominalizations function as Themes.
In (i), the understood complement subject is Alex, the Goal,
-
while in (ii), the understood complement subject is Mary,
the Source.
As is well known, the subject of the complement
of permit is controlled by the Goal of permit, and the subject of the complement of promise is controlled by the
Source of promise.
What appears to be needed, then, to
correctly account for the control relationships in (i),and
(ii) is a way of matdhing up thematic relations within the
nominalizations with thematic relations in the sentence.
Higgins (1973: 182-3) formulates the needed matching
principde as follows: "To each of the understood noun
phrases of the noun phrase whose head is the nominalization
there must correspond a noun phrase in the sentence which
bears the same thematic relation to the verb as that noun
phrase bears to the nominalization."
By this principle, the
complement subject in (i), which bears the thematic relation
of Goal to permission is correctly matched with Alex, which
bears the thematic relation of Goal to give.
Similarly for
(ii), the complement subject, which bears the thematic relation of Source to promise, is matched up with Mary, which
bears the thematic relation of Source to give.
121
In examples (8) and (9) in the text above, the manipulator of the delkbed object is matched with the manipulator of the antecedent of the deleted object.
16.
The control rule for the subject of infinitival rela-
tive clauses works in essentially the same way as the control rule for infinitival purpose clauses.
That is, the
subject NP of the relative clause, which exercises control
over the NP deleted under identity with the head, is matched
with the NP exercising control over the whole NP in the
matrix sentence.
Thus, the same principle is at work.
An incidental interesting fact about the matching
within a nominalization with the themaof themati~.~relations
tic relations of the sentence is that, is some instances,
there seems to be no single controller of the complement,
and there is, in fact, some problem in determining what the
interpretation of the complement subject is.
consider,
e.g=,
i.
permission to leave was transferred from John to
Bill
It is unclear what the understood subject of to leave is.
Observe how this works with respect to reflexivization:
ii.
permission to indulge himself was transferred from
John to Bill
122
iii. *permission to indulge oneself was transferred
from John to Bill
From,: these examplqs, we may conclude that the subject is not
interpreted as Unspecified.
But consider what happens if
one of the persons in the examples is female:
iv. *permission to indulge himself was transferred
from John to Agnes
v. "permission to indulge himself was transferred
from Agnes to John
vi. "permission to indulge herself was transferred
from John to Agnes
vii. "permission to indulge herself was transferred
from Agnes to John
The sentences are also out if oneself replaces himself/hera
self.
CHAPTER 111
A DIGRESSION ON GERUNDIVE FOR-PHWSES
1.
The prior-posterior distinction.
Consider now for-phrases whose objects designate a qual-
ity or attribute of someone or something.
Contrast:
1. I designed the car for endurance at highA,speeds
2. I bought the car for its endurance at high speeds
-
In both (1) and (2), the object of for designates an attribute of the car, but there is a crucial difference:
In (I), the property of have endurance at high speeds is
a result of the action of designing.
This is because design
designates a creative action; its object is understood as
being created or constructed by the action.
By contrast,
the cause-effect relationship between the matrix verb and
the for-phrase is reversed in (1). The buying of the car is
motivated by its quality of enduring at high speeds.
The
matrix predicate describes a subjective reaction to the attribute of the car in (2), and the car's endurance at high
speeds is understood to be semantically prior to the act of
buying the car.
In the presence of a motivational for-phrase if the
kind in ( 2 ) , the main verb describes some appropriate reaction inspired by the attribute of the object character-
ized in the for-phrase.
This description seems to do for
cases like (2), where there is a Theme-Goal relationship
between the object and the subject
--
i.e., with verbs
of volitional acquisition like buy, take, and get and also
with verbs of covetous emotions like want,
lave, and admire.
- But the semantic relationship between the for-phrase and
the matrix predicate is much more complicated, as can be
seen If we contrast (2) with
3.
I sold the car for its endurance at high speeds
for-phrase is understood as a facilitating
In (3), the factor in the transaction; the speaker is understood to be
using the car's attribute of endiiring at high speeds to
effect the sale.
Nevertheless, in (3), as in (2), the for-
phrase is understood to be-semantically prior to the matrix
P"
action. 1
There are some facts which may be taken to indicate a
structural difference between (1) and (2), although conclusive evidence seems hard to find.
We suggested in an
earlier discussion that resultative for-phrases like the
one in (1) are sister constituents of the direct object.
Verbs which take both purpose phrases and resultative phrases,
like build, cannot have them co-occurii~g:
4. a.
I built this car for the race next Tuesday
125
4. b.
I built this car for endurance at high speeds
c. *I built this car for the race next:Tuesday
for
endurance at high speeds
*I built this car for endurance at high speeds for
the race next Tuesday
This situation is explicable if we assume that verbs like
for-phrase which, depending on context, lexibuild take a cal content, etc., can have a purpose or resultative interpretation.
Thus, only one for-phrase complement is possible,
and its interpretation varies.
for-phrase
By contrast, the -
in (2) is compatible with the presence of a purpose phrase,
the principal restriction being that the purpose phrase must
precede the motivational phrase:
5.
I bought this car for Sunday drives for its
endurance at high speeds
*I bought this car for its endurance at high :
speeds for Sunday drives
This restriction indicates (a la Williams) that the motivational phrase is generated higher in the matrix tree than
the purpose phrase.
The prior-posterior distinction in the semantic rela-
-
tionship between for-phrases and the matrix predicate seems
to be of some~importance. (See, e.g., Chapter I1 of
Bresnan (1972). )
I would like to consider how this distinc-
126
tion interacts with the ability of for to take an infinitive
phrase or gerund phrase object in various contexts.
To start with, we must point out a crucial distinction in the interpretation of gerunds.
Wasow and Roeper
(1971) draw the distinction between nominal gerunds, which
have the internal structure of N P 1 s , and verbal gerunds,
which have the internal structure of sentences.
(Cp. John's
hunting of the snark and John's hunting the snark.)
They
correlate obligatory subject control with verbal gerunds and
absence of subject control with nominal gerunds, and demonstrate how this correlation can be explained within the
lexicalist framework on the basis of the fact that S 1 s but
not N P 1 s have an obligatory subject position.
The absence
of a subject in verbal gerunds can be effected only through
Equi-NP deletion.
(See their paper for discussion.)
Later on in their paper, Wasow and Roeper (henceforth,
W
& R)
discuss some counter-examples to their hypothesis.
One interesting class of cases involves yerbal gerunds
whose missing subjects are not controlled by any NP in the
matrix sentence.
Examples they give are as follows
p. 12 of their paper):
6. I disapprove of shouting loudly
7. the law forbids shooting deer
(See
127
Following a suggestion of Postal's, W
&
R suggest that in
such examples, the underlying subject of the gerund is the
one, which is deleted by an ad hoc transformgeneric pronoun ation:
8.
I disapprove of one's shouting loudly
9.
the law forbids one's shooting deer
one-deletion rule converts (8) and (9) to (6) and (7)
The respectively.
It seems to me that there are cases where such a solution will not work.
Consider examples like the follawing:
10:-aeqBilltalked to the boys about his seeing pink
elephants after excessive drinking
b. Bill talked to the boys about their seeing pink
elephants after excessive drinking
c. Bill talked to the boys about one's seeing pink
elephants after excessive drinking
11.
Bill talked to the boys about seeing pink elephants
after excessive drinking
The gerunds in (lo), with overt subjects, seem to me to have
a factive interpretation; they designate actual states of
affairs (I do not think,however, that they are precisely
paraphrased by the fact that S construction.)
Notice now that (11) seems to be three ways ambiguous:
128
t h e u n d e r s t o o d s u b j e c t of t h e g e r u n d c a n be B i l l o r t h e y b o y s
o r n e i t h e r of t h e s e .
~ h u s ,i t seems t h a t t h e u n d e r s t o o d sub-
j e c t of t h e g e r u n d i n (11) i s i d e n t i f i e d by some r u l e of
p r o n o m i n a l i z a t i o n , and we c a n p o s i t PRO o r
A
a s the subject
o f t h e g e r u n d , a t l e a s t f o r t h e f i r s t two r e a d i n g s where
t h e g e r u n d s u b j e c t i s c o n t r o l l e d by one o f t h e m a t r i x N P ' s .
For t h e t h i r d r e a d i n g ( w i t h t h e "unspecified" s u b j e c t r e a d i n g ) ,
W
would p o s i t o n e a s t h e u n d e r l y i n g s u b j e c t of t h e g e r -
& R
und, i . e . ,
il
p1
t h e y would d e r i v e (11) on t h i s t h i r d r e a d i n g
from ( 1 0 c ) t h r o u g h t h e d e l e t i o n o f o n e .
-
Such a d e r i v a t i o n i s d u b i o u s t o my mind.
( 1 0 c ) and
(11) on t h e t h i r d r e a d i n g a r e n o t synonymous.
A s I have
pointed o u t , t h e gerund i n (10c) has a f a c t i v e i n t e r p r e t a t i o n
and t h i s i s l a c k i n g i n (11) on t h e t h i r d r e a d i n g , where
t h e g e r u n d p h r a s e seems t o d e s c r i b e some h y p o t h e t i c a l e x p e r ience.
Compare, f u r t h e r , examples l i k e ,
1 2 . t h e d o c t o r s w e r e unhappy a b o u t o n e ' s h a v i n g g o t t e n
p r e g n a n t on t h e i r p i l l s o o f t e n
1 3 . t h e d o c t o r s w e r e unhappy a b o u t h a v i h g
n a n t on t h e i r
pill
g o t t e n preg-
so often
The s u b j e c t o f t h e g e r u n d i n ( 1 2 ) c a n n o t d e l e t e t o g i v e
( 1 3 ) ; ( 1 3 ) h a s no g e n e r i c r e a d i n g f o r i t s s t h j e c t .
The
the
s u b j e c t o f t h e g e r u n d i s o b a i g a t o r i l y c o n t r o l l e d by L
doctors.
Thus, W
& R's
t r e a t m e n t of t h e s e counterexamples
129
to their proposals is flawed by-he
utibn of verbal gerunds with
one
fact that the distrib-
as their subject differs
from the distribution of verbal gerunds with uncontrolled
subjects.
It is far from clear how to state the context
in which their rule would be applicable.
It is conceiv-
one-deleable that we can impose conditions on the ad hoc tion rule to accouht for the discrepancies, but, even so,
such an analysis fails to take account of the differences
between the interpretation of the two types of verbal gerunds.
Consider
one further pair of examples:
14. Jonathan contemplated having only 30 years to live
in Saudi Arabia
15. Jonathan contemplated one's having: only 30 years
to live in Saudi Arabia
(14) is an ambiguous sentence:
On one reading, the subj-
ject of the gerund is understood to be Jonathan, and it means
roughly that Jonathan contemplated that he would have only
30 years to live in Saudi Arabia.
On the second reading,
there is no. controller for the subject of the gerund, and
we understand that Jonathan is abstractly contemplating the
hypothetical situation of having only 30 years to live in
Saudi Arabia.
Note that this reading is not equivalent to
the reading of (15).
In (15), Jonathan is contemplating a
factual situation, viz. that one has only 30 years to live
in Saudi Arabia.
This is quite distinct from the second
reading of (14).
It seems to me that a more viable approach to this
pgoblem is to posit two structurally distinct types of verbal gerunds.
One type, as W
&
R suggest, is verbal gerunds
which have the internal structure of sentences.
The other
type is gerunds which have no lexically specified subject and
no subject controller and a non-factive interpretation,
and which have the internal structure of VP's.
In other
words, verbal gerunds can have either the structures of S's
dominated by NP's or the structure of VP's dominated by
NP's.
Of course, the full elaboration of this theory
would have to account for the distributional differences
between the two types of gerunds.
For example, the permis-
sibility of the VP-gerunds correlates with certain possibilities in the interpretation of nominal gerunds.
In (16), a
VP-interpretation of the gerund is allowed, while in (17) it
is not:
16. I was thinking about hunting the snark
17. I was glad about hunting the snark
In (17), the subject of the gerund is obligatorily understood as I; in (16), on the other hand, the gerund can optionally have a subjectless interpretation.
the following:
Consider now
131
18.
I was thinking about the hpnting of the snark
19.
I was glad about the hunting of the snark
The nominalization in (18) can be understood to designate either a hypothetical endeavor or an actual event;
only the latter interpretation is possible in (19).
I have raised these issues here because I wish to dis-
tinguish the interpretation of gerunds as objects of purfor-phrases from their interpretation as objects of
pose for-phrases.
"upper" motivational -
It seems that only nom-
inal gerunds and VP-verbal gerunds can serve as objects in
-
a purpose for-phrase.
20.
Consider the following paradigm:
I bought this knife for the hunting of the snark
(*with)
21.
I bought this knife for hunting the snark with
22.
I bought this knife for hunting the snakk,,
23. *I bought this knife for Bill's hunting the snakk
with
24. *I bought this knife for Bill's hunting the snark
25.
I bought this knife for Bill's hunting (of the
snark) (*with)
(21) and (22) show that verbal gerunds can serve as objects
to the purpose for-phrase, with the gerund phrase optionally being a semantically open expression; both (21) and (22)
are good.
This shows gerund phrases to,be different from
infinitive phrases as objects of a purpose for-phrase, since
infinitive phrases must, in these circumstances, be open
sentences.2
(23) and (24) show that gerund phrases with specified
subjects cannot serve as objects of a purpose phrase independently of the question of open sentences.
will be brought up again in the next chapter.)
(This fact
I suggest
that the anomaly of these examples is a result of the incompatibility of the factive element in the interpretation of
forsentential gerunds with the intentional meaning of the phrase.
Non-sentential verbal gerunds (i.e., VP-verbal
gerunds), which lack this factive element in their interpretation, are not in conflict with the intentional meaning
for-phrase and, hence, can serve as objects of the
of the purpose phrase.
~ o m i n a lgerunds are compatible objects of
purpose phrases, but, as (20) and (25) show, they cannot
be interpreted as open expressions.
Now, when a for-phrase is associated with a matrix
predicate which describes a reaction to a purpose or thing
such that the for-phrqse designates the reason for the reaction, a gerund phrase is a possible object of for, but
an infinitive-:phrase is not.
Consider, e.g., the verbs
blame and admire and the adjective proud:
26.
Bill blamed John for leaving the garden unattended
Bill admired John for having courage in the face
of doom
Bill was proud of John for rebelling against the
Establishment
*Bill blamed John to leave the garden unattended
*Bill admired John -to have courage in the face of
doom
*Bill was proud of John to rebel against the Establishment
for-phrases in (26) are understoad as semantically prior
The to the matrix predicates:
John's blaming Bill is predicated
on the garden's having been left unattended; Bill's admiration for John is based on John's show of courage; Bill's
pride in John is based on John's rebellion.
This is what I
mean by calling these predicates reactive; the cause-effect
chain leads from the for-phrase to the matrix predicate.
Notice that the gerunds here are interpreted as describing
actual states of affairs and that the gerund subjects are
controlled by the objects of the matrix predicates.
Their
interpretation is, thus, substantially different from the
interpretation of the gerund objects of purpose phrases.
As the*-examplesof (27) indicate, infinitive phrases are
unacceptable in this context; in characterizing motivations
for what goes on in the matrix predicate, infinitive phrases
..
.
I
8 '
Ill1 I
,
47'1
. , I<l8,.
8
I
" .I
1 ill
I
.
-
,
I
134
can only represent intentions and are understood as semantically posterior to the matrix predicate.
In other words
the cause-effect chain proceeds in the opposite direction,
-
from the matrix predicate to the for-phrase.
Thus, in a
sentence like
28. Bill blamed John for leaving the garden unattended
to throw any suspicion off himself
for-phrases represent, in some sense, motivations for
both the action of blaming, but whereas the act of blaming John
is predicated on the garden being left unattended, the removal of suspicion from Bill is predicated on the act of
blaming.
for-phrase is semantically prior
Thus, the first -
tothe act of blaming, while the act of blaming is in turn
for-phrase (i.e., the rationale clause).
prior to the second Put another way, the fpr-phrase complement to a reactive
predicate like blame is interpreted as motivational and nonintentional, actualized; it repreGents an actual event or
state of affairs.
This precludes infinitive phrases as
objects of such phrases, because of their intentional,
non-actualized interpretation.
There are some cases, however, in which an infinitive
phrase can have a non-intentional, actualized interpretation, as in
29. I am glad to be feeling well again
30. I am proud to present you with this award
In these examples, the infintive phrase represents the
"Goal" of the reaction, much as John is the "Goal" of the
reaction in the examples of (26). The infinitive phrases
in (29) and (30) arc motivational, in the sense that they
respresent the situations that call forth the emotional
reaction of gladness or pride.
Bresnan (1972~79)cites
some further cases where both gerund for-phrases and infinitive phrases have non-intentional interpretations and are
understood as motivations for moral judgments.
However,
in most of these cases, as in (29) and (SO), it is unlikely that the infinitive phrases are objects of the preposition for:
31. I am glad of/*for that
32. I am proud of/*for that
In general, infinitive phrase objects of for have a nonactualized, intentional interpretation, and are understood
as semantically posterior to the matrix predicate.
Gerund
fdr have an actualized interpretation and
phrase objects of -
are understood as semantically prior to the matrix predicate.
Thus, the distributions of infinitive phrases and gerund
phrases as objects of for seem to be nearly complementary,
with the exception of the non-factive gerunds which can
136
for-phrases as well as infinioccur as objects of purpose tives.3
We have seen that sentential gerunds, which are factive as object of for4, are unacceptable as objects of
for-phrases.
purpose -
They are, likewise, unacceptable as
for-phrase comobjects of objective phrases and of the plements of psychological predicates, which have an intentional interpretation:
33.
Jon sent Fred to get his books
*John sent Fred for getting his books
34.
Jim was eager to see his first snowstorm
*Jim was eager for seeing his first snowstorm
35.
Amanda was anxious to get away
*Amanda was anxious for getting away
One further demonstration of the differences in the
interpretation ~f infinitives and gerunds as objects t~
for involves nominals taking for-phrase complements like
-
-
permission, grant, chance, need, excuse, desire, etc. Permission differs from its verb counterpart permit in several
ways:
The nominalization takes no indirect object, while
the verb does:
36.
permission to leave-.cameyesterday
"permission to/of John to leave came yesterday
they permitted John to leave yesterday
137
The verb takes a direct object and an optional indirect
for-phrase.
object but no for-phrase
37,
The nominalization takes only a
and not a direct object:
they permit visits
they permit John visits
*they permit (John) for visits
permission for a visit/*of a visit
Also, while there is optional control of the subject of an
infinitive complement to the nominalization permission, the
subject of an infinitive complement to the verb permit is
obligatorily contrdlled by the Goal:
38.
they gave the US permission to withdraw
they gave the US permission for the troops to
withdraw
they permitted the U S to withdraw
permitted the U S for the troops to withdraw
The verb excuse also differs crucially from its nominalization excuse.
Both take for-phrase complements: an excuse
for my bad behavior, she excused me for my bad behavior.
However, the nominalization can take both infinittval and
and gerundive for-phrase complements while the verb can
take only gerundive for-phrase complements:
excuse to leave,
excuse for leaving, *she excused me to leave, she excused me
138
for leaving.
Also, the verb can take a simple direct object
while the nominalization cannot: she excused the error,
she excused my leaving, *excuse of the error, *excuse of my
leaving.
Now, the interpretation of the nominalization is diffor-phrase from when it takes
ferent when it takes a gerund for-phrase.
an infinitival -
The infinitive complement repre-
sents an intention, and the nominalization excuse to X is
interpreted, roughly, as some sort of justification that is
required before the intention can be carried out.
On the
other hand, the gerund phrase represents some already existing state of affairs or an action which has already been accomplished, or is in the process of being accomplished.
Notice, for example, that a nominalization like excuse for
being tall or excuse for having no checking account is possible while *excuse to be tall is not permitted.
The latter
nominalization is excluded because of the intentional meaning
the infinitive.
Intentionality implies volitionality
and control over an action, and, thus, to be tall does not
qualify as a proper complement.
On the other hand, a
gerund complement is not interpreted as intentional with
respect to excuse.
Excuse for V-ing means justification for
something ongoing or something which already is the case.
There are other aspects of the difference between the
two.
In comparing the phrases John's excuse for leaving and
John's excuse to leave, note that John is understood as the
Agent-Source of excuse only in the first case.
That is, in
John's excuse for leaving, John is understood to be providing
the justification for his actions.
In John's excuse to leave,
John is simply the Goal of excuse; excuse here denotes-some external factor or circumstance which allows John to
P
realize his intention.
We can see this difference in com-
parison of the following sentences:
39. a.
John's excuse for leaving wasn't exactly brilliant
b. *John's excuse to leave wasn't exactly brilliant
c.
John's excuse to leave came when his wife phoned
The difference between the examples of (39a) and (39b) is
analogous to the difference between:
40. a.
John's attempt wasn't exactly brilliant
b. *John's death wasn't exactly brilliant
In this contexk, only a nominalization with an agent subject is possible.
~ o t i c e ,in addition, that excuse here
must be interpreted as a communicable product of John's
efforts.
41.
Compare:
John gave the hostehs an excuse for leaving early
the candidate gave his committee an attempt to
solve the mind-body problem as a dissertation
140
41. *the general gave ~ i x o nan attack on a Vietnamese
village as a justification for his promotion
The phrase an attack on a Vietnamese village
cannot be
interpreted as a communicable product, and the last sentence of (41) is excluded.
Consider now the difference between the follawing:
42.
I gave the hostess an excuse for leaving the
party
43.
I gave the hostess an excuse to leave the party
The subject of leave in (42) is I, while the subject of
leave in (43) is the hostegs.
Thus, excuse with a gerund
for-phrase marks its Agent-Source as controller of the gerund
subject, and this is matched with the Agent-Source of the
matrix sentence by Higgin's matching principle.
On the
other hand, excuse with an infinitival for-phrase complement
marks its Goal NP as controller of the complement subject,
and this is matched with the Goal NP of the matrix sentence.
In this way, the difference between (42) and (43) seems
parallel to thekrdifference between (44) and (45):
44.
I gave the hostess a promise to leave the party
45.
I gave the hostess permission to leave the party
The parallel is far from exact, however, because (42) is
actually ambiguous: there is a reading on which the subject
of leave is understood to be the hostess; this is brought
out more clearly in sentences like
My presence gave the hpstess an excuse for leav-
46.
ing the party
I (my presence) gave the hostess an excuse for
ordering me off the premises
It seems to me that on this second reading of (42), the
sentence is paraphrasable by (47a):
47. a. I gave the hostess a means of excusing herself
<
<
I
from the .party
b.*I gave the hostess a means of excusing myself
from the party
c. My presence gave the hostess a means of excusing
herself from the party
Notice, incidentally, that the sentence I excused myself for
drinking heavily last night is ambiguous.
On one reading,
it is comparable to I excused John for drinking heavily last
night, with myself simply replacing John.
On the second
reading, myself is a bound occurrence of the reflexive.
There is an expression excuse oneself which has properties
different from the simple verb excuse.
Principal among the
differences is the fact that excuse oneself
to-phrase:
-
can take a
I excused myself to Mary for drinking heavily
48.
last night
*I excused John to Mary for drinking heavily last
night
One pair of sentences which further demonstrates the difference in the interpretation of the nominalization excuse
with each of its two possible sentential complements is as
follows:
49.
my excuse to leave was my wife's telephone call
my excuse for leaving was my wife's telephone
call
Tn the first sentence, my wife's telephone call is understood
as the factor which allows me to leave; in the second sentence, it is understood as the substance of the excuse I
offer for the fact that I am leaving or have left.
The noun reason is quite similar to excuse, but there
are interesting differences.
Like excuse, reason can take
-
for-phrase
either a gerund for-phrase or an infinitive complement: reason to leave the party, reason for leaving the
party.
-
Unlike excuse, however, the
infinitive complement
does not have an intentional interpretation.
.
Compare, e.g.,
reason to have left, *excuse to have left, reason to hate
Nixon, *excuse to hate Nixon.
Nevertheless, the infinitive
QI
complement must represent something which is controllable:
reason for being tall, *reason to be tall.
As with excu.se, the gerund phrase is interpreted as
designating an existing state of affairs, and the phrase
reason for V-ing is descriptive of an underlying causal
factor.
With an infinitive phrase complement, the head
noun reason-indicates a causal factor which is semantically
prior to the instantiation of the infinitive phrase.
It
is interesting to note that reason with an infinitive phrase
complement differs from excuse in not being able to take a
definite determiner:
50.
the excuse to leave came when his answering service telephoned
his excuse to leave was his answering service's
telephone call
* t h e reason to leave came when his answering ser-
vice telephoned
*his reason to leave was his answering service's
telephone call
Consider now the analogs to ( 4 2 ) and ( 4 3 ) .
51.
I gave the hostess a reason for leaving the party
52.
I gave the hostess a reason to leave the party
144
A s with
( 4 3 ) , t h e s u b j e c t of t h e i n f i n i t i v e p h r a s e i s un-
ambiguously t h e h o s t e s s .
And, a s w i t h ( 4 2 ) , t h e s u b j e c t
I or the hostess.
of t h e gerund p h r a s e i s e i t h e r -
However,
t h e r e i s a d i f f e r e n c e : I n (511, t h e m a t r i x p r e d i c a t e cann o t be an i n t e n t i o n a l a c t i o n on t h e p a r t of t h e m a t r i x
I i f t h e h o s t e s s i s u n d e r s t o o d a s s u b j e c t of t h e
subject -
infinitive.
53.
Thus,
(53) i s f i n e :
my p r e s e n c e gave t h e h o s t e s s a r e a s o n f o r l e a v i n g
the party
b u t compare t h e f o l l o w i n g :
54.
B i l l had t h e n e r v e t o g i v e t h e h o s t e s s a n excuse
f o r taking h i s leave
B i l l had t h e n e r v e t o g i v e t h e h o s t e s s an excuse
f o r taking her leave
B i l l had t h e n e r v e t o g i v e t h e h o s t e s s a r e a s o n
for taking h i s leave
* B i l l had t h e n e r v e t o g i v e t h e h o s t e s s a r e a s o n
f o r taking her leave
I t h i n k t h i s h a s t o do w i t h t h e f a c t t h a t an excuse b u t n o t
a r e a s o n i s c o n t r o l l a b l e i n t h e s e n s e t h a t t h e s u b j e c t of t h e
gerund p h r a s e h a s t h e o p t i o n of a p p l y i n g i t :
55.
the hostess has an excuse for leaving, if she wants
*the hostess has a reason for leaving, if she wants
Compare the following:
56.
the hostess has a reason to leave,-'ifshe wants
to/*it
the hostess has an excuse to leave if she wants
to/*it
In sum, we find that gerund phrases in these contexts are
interpreted as factive and semantically prior to the nominal
head; infinitive phrases are interpreted as semantically
posterior to the nominal head.
Some nouns can take comple-
ments of only one or the other type.
Examples of nouns
taking only gerundive complements are the following: motive
(motive for murder, motive for murdering someone, *motive to
murder someone), justification (justification for that act,
justification for acting that way, *iustification to act that
way) (cf. I am justified to act that way/*for acting that
way/ *for that action.
The infinitive phrase here is not a
for-phrase.), apologies (apologies for his bad behavior,
apologies for behaving so badly, *apologies to behave so
badly.)
Examples of nouns taking only infinitive complements
are: urge (urge for an ice cream cone,
.,. urge to have an ice
146
cream cone, *urge for having an ice cream cone), desire
(desire for that, desire to do that, *desire for doing that)
and cause (cause for alarm, cause to.cry/*for crying).
2.
Observations on control in higher-generated for-phrases.
We have discussed before mechanisms by which the sub-
jects of rationale clauses and lower-generated infinitival
for-phrases are uniquely
-
determined; the subject. of a
rationale clause, when not overtly present, is controlled
by the matrix subject.
For clauses of the type discussed
in Appendix A of the last chapter, the subject is controlled
by that NP which the main verb designates as Agent of the
complement, a la Jackendoff.
1.
Thus, in examples like
John paid me to waste my time
the subject of the infinitive phrase is the matrix subject John, if the infinitive phrase is interpreted as a
rationale clause, and me
- otherwise.
Let us compare sub-
ject control for gerundive for-phrases. With verbs like
blame and admire, the object NP is designated as controller
of the gerund phrase, i.e., as the NP which is understood
as representing the individual responsible for the event
described in the gerund phrase:
147
2.
I blamed Johni for - letting the cat out of the
bag
I admire Johni for - not speaking until he is
spoken to
Actually, there are differences between these two verbs.
The
for-phrase complement designate
verb admire requires that the something specifically about its object
--
some quality or
action or activity for which the object bears direct responsibility.
With blame, on the other hand, the object can
have direet or indirect responsibility.
3.
Consider:
I admire my parents for an inability to procrastinate
4.
I blame my parents for an inability to procrastinate
In ( 3 ) , we are talking only of an inability on the part of
parents to procrastinate.
In (4), the inability to pro-
crastinate can be attkibuted either to the parents or to
the speaker:
5.
I admire my parents for their inability to procrastinate
*I admire my parents for my inability to procras-
tinate
6.
I blame my parents for their inability to procras-
tinate
6.
I hlame my parents for my inability to procras-
tinate
With gerund objects of for, this contrast between admire and
blame is reflected in obligatory vs. optional subject control.
7.
I admire Gopali for - daring to question the
Guru
?I admire Gopali for hisi dating to question the
Guru
*I admire Gopal for Sanchi's daring to question
the Guru
*I admire Gopal for - daring to question the
Guru
8.
you can blame Johni for
not getting out on
time
youi can blame John for - not getting out on
time
you can blame Johni for hisi not getting out on
time
you can blame John for your/Billls not getting out
on time
Nevertheless, even with blame, the matrix object is interpreted as Agent of the for-phrase.
Thus, when the gerund
149
describes an Agent-controlled action, subject control must
go to the object of blame.
(Compare the scream-cases dis-
cussed in Chapter 11, Appendix A.)
we blamed Johni for
killing the .pig
*wei blamed John for
killing -the pig
9.
(This is, of course, exclusive of cases like, we blamed John
for hurting ourselves, which have only the well-known
"accidental" reading.)
The verb punish works like the verb admire.
The sub-
ject of the gerund phrase is obligaforily controlled by the
object of punish.
10.
she punished Maxi for
-i
making a mess
?she punished Maxi for hisi making a mess
*she punished Max for the dog's making a mess
The verbs compensate and pay are interesting in that they
allow ambiguous control of the gerund phrase subject.
Thus,
there is a contrast between the case where pay takes a
lower for-phrase complement, where there is a matrix
-
Agent, and the case where pay takes a gerund for-phrase.
Consider
11.
Max paid me to waste my time
Max paid me for wasting my time
12.
Max compensated me for wasting my time
The control of the complement subject is unambiguously assigned to me
- on the lower for-phrase reading of the first
sentence of (ll), but in the second sentence, the subject
-
of wasting my time can be either Max or me.
-
It is possible
that the sentence has different structures on each reading,
but I am not positive about the status of the relevant
data.
If sentences like, Max paid me for fixing his gold
watch for wasting my time or for wasting my time, Max paid
me for.fixing his gold watch are grammatical, then we can
argue that the for-phrase is generated at a higher level in
the matrix tree in the case where the subject of the gerund
phrase is understood to be Max.
-
It is qaite possible that
more than two levels of structure will be required to generate all the complement types.
.
(See Williams (1971) )
In
the case of compensate, the analogous examples do not seem
grammatical; for-phrases with gerund objects cannot be
stacked: *Max compensated me for making the cookies for
wasting my time; *for wasting my time, Max compensated me for
making the cookies. )
Like rationale for-phrases with infinitive objects,
for-phrases can function semantically as complegerundive merits to modals.
Compare:
13.
the soldier punished his son for falling asleep
on guard duty
14.
the soldier had to punish his son for falling
asleep on guard duty
The subject of the gerund phrase in (13) is uniquely interpreted as his son, in line with our previous observations.
However, the complement subject in (14) is ambiguous; it can
On the
be understood as either the soldier or his son.
reading where the soldier is understood as the subject of
the gerund phrase, the for-phrase is interpreted as a
complement to the modal.
The modals which--allowfor gerund
for-phrase complements are those which involve a sense of
obligation: must, should, ought to, have to. The gerund forphrase describes what motivates the obligation on the part,
of the matrix subject.
Thus, in (141, falling asleep on
guard duty incurs an obligation on the part of the soldier
to punish his son.
for-phrase complements to
Observe that -
formodals are generated higher in the matrix tree than the phrase complement to punish:
15.
the soldier had to punish his son for playing
hookey for falling asleep on guard duty
forNotice that it is impossible to interpret the first phrase as the complement to the modal, which would allow
the soldier to be understood as the subject of playing
I
'
--
------*&a.
>--
-- ., .--, T
I
'
,
1, Ti
-
-
- -
--
152
hookey.
Also, when an infinitival for-phrase is complement
to the modal, the matrix subject cannot, of course, be the
forcontroller of the gerund phrase subject of the first phrase:
16.
Judge Carswell had to punish the lady witness for
perjuring herself
Judge Carswell had to punish the lady witness for
perjuring himself
Judge Carswell had to punish the lady witness for
perjuring herself to stay on the bench
*Judge Carswell had to punish the lady witness for
perjuring himself to stay on the bench
The
only way the last sentence of (16) can be grammatical
is for the infinitive phrase to be interpreted as a rationale
clause within the gerund phrase.
Finally, observe that with the verb get, the Goal NP
is interpreted as understood subject of the gerund phrase,
as in
17.
Bill got the brush for cleaning the room so well
Bill got Mary the brush for cleaning the room so
well
Similarly, with give,buy, etc.:
153
18.
B i l l gave Mary t h e b r u s h f o r c l e a n i n g h i s room
so well
B i l l bought Mary t h e b r u s h f o r c l e a n i n g h i s room
so well
Ih ( 1 7 ) and ( 1 8 ) , t h e u n d e r l i n e d NP ' s a r e t h e Goals, which
a r e u n d e r s t o o d a s t h e s u b j e c t s of t h e gerund p h r a s e s .
To
be c l e a r , t h e Goal i s u n d e r s t o o d s u b j e c t of t h e gerund p h r a s e ,
when t h e m a t r i x p r e d i c a t e i s i n t e r p r e t e d a s a rewarding o r
for-phrase.
p u n i t i v e r e a c t i o n t o what i s d e s c r i b e d i n t h e buy, even when it i s
N o t i c e t h a t i n t h i s s e n s e , t h e Agent of t h e m a t r i x Goal, c a n n o t c o n t r o l t h e gerund p h r a s e s u b j e c t :
1 9 . # B i l l bought t h e b r u s h f o r c l e a n i n g h i s room s o
well
Compare t h i s w i t h :
20.
B i l l r e c e i v e d t h e b r u s h f o r c l e a n i n g h i s room s o
well
where B i l l i s Goal b u t n o t Agent.
However, t h e gerund for-
p h r a s e can be u n d e r s t o o d a s t h e m o t i v a t i o n f o r some compensat o r y a c t i o n on t h e p a r t of t h e m a t r i x Agent, i n which c a s e
t h e m a t r i x Agent s u b j e c t d o e s c o n t r o l t h e s u b j e c t of t h e
gerund p h r a s e , a s i n
21.
Bill gave Mary an ice cream cone for having
wasted her time the day before
Bill bought Mary an ice cream cone for having
wasted her time the day before
(I think the control is ambiguous in these examples.)
The
verb compensate allows for both construals of the motivational phrase (a rewarding for action on the part of the
matrix Goal or compensation for some debt-incurring action
on the part of the matrix Agent); hence, control is ambiguous.
Once again, it is possible that the ambiguity of the
sentences of (21) is structural, but clear evidence seems
quite hard to find.
-
What all this shows is that, in the case of gerund forphrases, the control of the complement subject is free but
subject to the influence of various aspects of the interfor-phrase and
pretation of the relationship between the the matrix predicate.
(At least from this point of view,
sentences like John hit Bill for stealing grapes (See Note
5 of the first chapter) present no real problem for Williams'
analysis. (Williams (1971))
The assignment of control of
the subject of the gerund phrase works essentially like
pronominalization, so that we would not expect it to be any
more sensitive to the relative position of the for-phrase
and the controlling NP than any other kind of pronominalization. )
As far as infinitival rationale clauses are concerned,
excluding now the case where they are complements to a modal
or to a conditional predicate, etc., the matrix predicate
is interpreted as a volitional action on the part of the
subject; the rationale clause describes the intention of
that action and is semantically posterior to the matrix
predicate; hence, the matrix subject is understood as an
Agent with respect to the rationale clause.
Thus, even
when the matrix predicate is passive, it has to be interpreted as subject to the volition of the matrix subject with
respect to the intention depicted in the rationale clause:
22. Sam was arrested in Dallas to make a name for
himsely .
Most passive examples are, I think, dubious.
(Note, inci-
dentally, that the possibility of a volitional interpretation
is an important difference between the construction seem
predicate and the construction seem
+
+
infinitive; for me,
the sentence, Sam seemed cool to throw suspicion off himself
acceptable, while Sam-seemed to be cool to throw:suspicion off himself is not.
Compare the two constructions as complements to volitional verbs like try or appear:
Sam tried to seem (*to be)
Sam appeared to seemp(*to be) cool.
It is possible, however, to find interesting cases
156
where the subject of the rationale clause is not understood
to be the matrix subject.
In examples like
23. Kennedy was killed by certain unknown agents to
prevent curtailment of the CIA's powers
Tom was put in the driver's seat to confuse the
police
the matrix subjects are not understood as Agents, and the
subject of the rationale clause seems to be the matrix
predicate itself:
24. Kennedy's being killed by certain unknown agents
prevented curtailment of the CIA's powers
Kennedy's being assasinated by unknown agents
was to prevent curtailment of the CIA,!s powers
Tom!s being put in the driver's seat confused the
police
Tom's being put in the driver's seat was to confuse the police
(The second sentence of (23) has, of course, an objective
clause-reading which we can ignore.)
Further, this kind
of situation arises also in cases where the rationale
clause is complement to a modal or a conditional predicate:
157
25. Oswald had to be arrested by the police to prevent
suspicion from arising in the minds of the people
it was necessary for Oswald to be arrested by the
police to prevent suspicion from arising in the
minds of the people
These facts need more thorough study.
FOOTNOTES
CHAPTER I11
1. All of this is not to mention contrasts like
i.
ii;
I bought the car for its bucket seats
I took the car apart for its bucket seats
In (i),
the bucket seats are understood as the attribute
of the car that attracted the speaker and prompted the purchase.
There is the possibility of understanding (i) in
such a way that we infer that the speaker is interested
only in the bucket seats, not in the whole car, and buys
the car so that he can have access to them.
This brings
us closer to the way (ii) must be understood; in (ii),
we understand that the action of taking the car apart is
aimed at getting the bucket seats.
Thus, while the object
of for in a sentence like (i) can be anything designating
a positive attribute, the motivating attribute in (ii) must
involve a concrete object.
iii.
Contrast:
I bought the car for its bright color
I bought the car for its smooth and easy ride
iv. *I took the car apart for its bright color
*I took the car apart for its smooth and easy ride
-
~ l of
l this, of course, goes to show that the factors influ-
r"
encing the interpretation in all these examples are numes-
159
ous and complicated in th8.way they interrelate and, for the
most part, not readily subject to formalization.
2.
Relevant judgments are not always clearcut.
With a verb
like buy, I judge semantically closed infinitive phrases as
unacceptable objects of the purpose phrase (i.e., as unacceptable purpose clauses).
That is, the rule of deletion or
interpretation which applies in purpose clauses seems to be
obligatory.
i.
ii.
:
I
bought this knife to hunt the snark with
?I bought this knife to hunt the snark
When the infinitive phrase has a lexically specified subject, the difference seems to me to be sharper:
iii.
I bought this knife for Karin to hunt the snark
with
iv.
*I bought this knife for Karin to hunt the snark
(iv) seems to me acceptable only with a rationale clause interpretation.
The matter is complicated by the fact that
the verb use uniquely accepts infinitive phrases of both
types :
v.
vi.
I used this knife to hunt the snark with
I used this knife to hunt the snark
160
I assume that the infinitive-phrase in (vi) is not strictly
speaking a purpose clause.
(It would be nice if I could
find a genuine ambiguity in the interpretation of non-clausal
for-phrase complements to use which would correlate with
this, but I haven't come up with anything convincing.)
Note too that when purpose clauses figure in copular sen-tences, the infinitive phrase must be "open":
vii.
this knife is to hunt snarks with
??this knife is to hunt the snark
this knife is for Bill to hunt the snark with
*this knife is for Bill to hunt the snark
I am inclined to take the most rigid position and claim that
purpose clauses must be open sentences, which leaves me with
the task of accounting for the difference between infinitive
phrase objects and gerund phrase objects.
(There is the
further difference that infinitival purpose clauses can have
lexically specified objects.)
Notice (whether or not this
has anything to do with the problem at hand) that there are
a number of additional differences between gerunds and infinitive phrases.
viii.
Compare, for instance:
I chose this knife to hunt with
I chose this knife for hunting with
I chose this fox to hunt
??I chose this for hunting
In these examples, it seems odd for the:object of choose to
control the ,abject of the gerund phrase, while there is no
such oddity in the case of the infinitive phrase.
Further,
consider the contrast between ready on its objective (as
opposed to subjective or psychological) interpretation, and
ripe:
ix.
the grass is ready for mowing
the grass is ready to mow
the grass is ripe for mowing
*the grass is ripe to mow
(See Note 14 of the last chapter.)
With object-deletion, the
infinitive phrase seems to carry an implication of some
external influence on the state of the grass.
This is seen
in the difference between the two ready examples.
Ripe,
precluding such an implication, since it describes an inherent state, does not allow an infinitive complement.
Actual-
ly, there are further complications, but such data indicate
that there are subtle differences in the interpretation of
gerunds and infinitive phrases in these contexts, and these
may ultimately be seen to prhvide an explanation for the
contrast between (21) and (22), on the one hand, and (i) and
(ii) on the other.
3.
There is an interesting argument to support our struc-
tural proposal for the gerunds that occur as objects of
purpose for-phrases.
We have seen that the Theme-controlled
NP in infinitival purpose clauses can be generated in sub-
ject position as well as object position:
i.
ii.
-
I bought the cati to catch mice with
I bought the cati
-i
to catch mice
We have also seen that a Theme-controlled NP can be generated in an object position in gerundive purpose phrases, as
in
iii.
I bought the cati for catching mice with
-
Notice, however, that in the following example, the cat is
not understood as the subject of the gerund:
iv.
I bought the cat for catching mice
Consider these further examples:
v.
Blind meni use seeing-eye dogs
-
to get them-
selves across the street
vi.
Blind men
j
use seeing-eye dogsi
-i
to get them
across the street
vii.
Blind men use seeing-eye dogs for getting themselves across the street
j
163
viii. *Blind meni use seeing-eye dogs for getting themi
across the street
The contrast between (vi) and (viii) shows that seeing-eye
dogs is not the subject of the gerund in (viii). This contrast is explicable if the infinitive phrases are analyzed
as sentences and the gerund phrases as VPrs: There is no
possibility of generating a Theme-controlled NP in subject
position in the gerund phrases because there is no subject
position.
Note that the reflexive
themselves which is coref-
erential with blind men presents no prablem here.
The
gerund phrase designates an actkvity which is engaged in
use,
by the NP which exercises control over the object of namely the matrix xubject.
Thus, the subject of the sen-
tence is designated as the Agent of the purpose phrase.
In
the case of purpose infinitives, this serves to determine
a coreference relationship between the subject of the infinitive-complement and the matrix Agent, but Agent-assignment
is independent of complement subject interpretation, as we
see from cases of indirect Agency.
By contrast, for (viii)
to be grammatical, there would have to be an empty subject
position in the gerund phrase to serve as coreferent of the
object of use, and the fact that (viii) is not grammatical
shows, therefore, that there simply is no subject position.
164
4.
Sentential gerunds do not have a factive interpretation
in all contexts.
In fact, the whole matter of the interpre-
tation of sentential gerunds is very complicated and needs
further study.
Let me just mention one aspect of the problem
which I find particularly fascinating. Consider the sentence
i.
Penny talked about quitting her job
(i) is ambiguous; the gerund can be interpreted as hypothetical or as descriptive of an actual state of affkirs;
iile., Penny can either be talking about the possibility or
the fact of quitting her job,
If the gerund has a specified
subject, only the factive interpretation is possible:
ii.
Penny talked about her/Sallyls quitting her job
Now consider the following:
iii.
Penny talked to Ray about quitting her job
iv.
Penny talked to Ray about quitting his job
The gerund in (iii) has either a hypothetical or factive
interpretation, while the gerund in (iv) has only a hypothetical interpretation.
This is shown if we add a past time
adverbial or the perfective aspect to bring out the.actua1ized interpretation of the gerund:
v.
Penny t a l k e d t o Ray a b o u t q u i t t i n g h e r j o b
yesterday
-
.
Penny t a l k e d t o Ray a b o u t having q u i t h e r job
(yesterday)
v i . *Penny t a l k e d t o Ray a b o u t q u i t t i n g h i s job p e s t e r day
*Penny t a l k e d , t o Ray a b o u t having q u i t h i s job
yesterday
( O f c o u r s e , t h e p e r f e c t i v e a s p e c t i s n o t i n g e n e r a l incompat-
i b l e w i t h t h e h y p o t h e t i c a l i n t e r p r e t a t i o n of t h e g e r u n d , a s
c a n be s e e n from examples l i k e , Penny t a l k e d t o Ray a b o u t
having gone t o c o l l e g e b e f o r e he would be f o r t y .
The ex-
c l u s i o n of t h e h y p o t h e t i c a l r e a d i n g w i t h t h e i n t r o d u c t i o n - ) o f
t h e p e r f e c t i v e a s p e c t depends on c o n t e x t , t h e p a r t i c u l a r
v e r b a t t h e head of t h e g e r u n d , and t h e p r e s e n c e of c e r t a i n
adverbial modifiers.)
Note f u r t h e r t h a t a f a c t i v e i n t e r p r e -
t a t i o n of t h e g e r u n d s i n ( v i ) i s p o s s i b l e i f we i n c l u d e
s p e c i f i e d s u b j e c t s f o r t h e gerunds:
vii.
Penny t a l k e d t o Ray a b o u t h i s q u i t t i n g h i s job
yesterday
Penny t a l k e d t o Ray a b o u t h i s having q u i t h i s job
yesterday
Penny t a l k e d t o Ray a b o u t t h e i r q u i t t i n g t h e i r
jobs yesterday
166
vii.
Penny t a l k e d t o Ray a b o u t t h e i r h a v i n g q u i t
t h e i r jobs yesterday
I t seems t o b e t h e c a s e , t h e n , t h a t i n c o n t e x t s l i k e ( v ) ,
t h e s u b j e c t of t h e g e r u n d i s o b l i g a t o r i l y c o n t r o l l e d by Penny
i f t h e gerund has a f a c t i v e i n t e r p r e t a t i o n .
On t h e hypo-
t h e t i c a l r e a d i n g of t h e g e r u n d , c o n t r o l o f t h e g e r u n d subj e c t i s f r e e , a s (iii) and ( i v ) show.
Ray c a n
-
I n f a c t , Penny and
" s h a r e " c o n t r o l on t h i s i n t e r p r e t a t i o n :
viii.
Penny t a l k e d t o Ray a b o u t q u i t t i n g t h e i r j o b s
( v i i i ) , o f c o u r s e , d o e s n o t have a £ a c t i v e - g e r u n d i n t e r p r e tation.
i x . *Penny t a l k e d t o Ray a b o u t h a v i n g q u i t t h e i r
jobs yesterday
T h e r e i s some i n d i c a t i o n t h a t t h e f a c t s we have j u s t n o t e d
a r e r e p r e s e n t a t i v e o f a more g e n e r a l phenomenon.
x.
Consider:
Ramsey C l a r k t a l k e d a b o u t a t r e k t h r o u g h t h e
j u n g l e t o Hanoi
I want t o f o c u s a t t e n t i o n on t h e i n t e r p r e t a t i o n o f t h i s sen-
t e n c e where t h e o b j e c t o f a b o u t i s u n d e r s t o o d a s d e s c r i p t i v e
of an a l r e a d y accomplished a c t i v i t y .
O n t h i s r e a d i n g , we
~ a ~ u n d e r s t a nt hd e - s e n t e n c e t o mean t h a t Ramsey C l a r k hims e l f made t h e t r e k .
T h e r e i s a n o t h e r i n t e r p r e t a t i o n , on
167
which the object of about reppesents a hypothetical, unrealized, intended action which Clark is understood to be
considering.
These interpretations seem parallel to the
interpretation of (i).
xi.
Consider now the following:
Ramsey Clark talked to reporters about a trek
through the jungle to Hanoi
On the hypothetical reading of the object of about, Ramsey
Clark can be understood to be proposing that he himself make
the trek, or thqt the reporters make the trek, or that he
and the reporters make the trek.
However, on the interpreta-
tion where the about-phrase object is understood as a fait
accompli, Ramsey Clark but not the reporters can be understood to have undertaken the trek.
This is again parallel
to the interpretation of the gerund phrases, as (v) and (vi)
show.
Incidentally, in (x) and (xi) on both readings, there
is also the possibility of interpreting the subject of trek
as unspecified.
his unspecified~:sub~ectinterpretation is
not possible for the gerund phrases discussed above, and
this contrast is precisely of the type analyzed by Wasow and
Roeper in the paper cited above.
That is, the possibility
of the unspecified subject interpretation in (x) and (xi)
is a function of the optionality of the subject position in
NP's as opposed to the obligatoriness of the subject position
in sentential gerunds.
We do, however, find sentences like,
Penny talked (to Qay) about quitting one's job.
But notice
that the interpretation of the gerund in this example is
distinct from the interpretation of the gerunds in (i),
( i , and ( v ) . The gerund here lacks the element of
possibility and intentionality khat is characteristic of
the earlier examples.
The sentence can be paraphrased
roughly as, Penny talked (to Ray) about the business/ the
matter of quitting one's job.
On the Postal-Wasow-Roeper
theory, we wsuZd derive the sentence from-,Penny talked
(to Ray) about one('s) quitting one's job.
However, as I
have suggested previously, such a derivation does not seem
in tune with the semantics of the alleged variants.
For
example, consider the factive element in the interpretation
.
of (xii)
xii.
Penny talked (to Ray) about one's usually quitting one's job in this place after only three weeks
*Penny talked about usually quitting one's job in
this place after only three weeks
Also, consider contrasts like, John talked about living high
on the hog,in Amsterdam
vs. John'talked
-high on the hog in Amsterdam.
about one's living
The subjectless gerunds in
these cases describe abstraot activities or experiences,
while the gerunds with one as their subject describe actual
169
s t a t e s of a f f a i r s .
Thus, t h e s e p a r t i c u l a r s u b j e c t l e s s g e r -
unds a r e i n a c l a s s by t h e m s e l v e s and a r e , on o u r a n a l y s i s ,
v e r b a l g e r u n d s w i t h t h e i n t e r n a l s t r u c t u r e o f VP's.
Observe
t h a t t h e c o n t r a s t between ( x i i ) and ( x i i i ) s t r o n g l y s u p p o r t s
t h i s analysis:
S i n c e u s u a l l y i s a s e n t e n t i a l adverb, it i s
p r e c l u d e d from t h e VP-verbal g e r u n d i n ( x i i i ) .
(As I re-
c a l l , s i m i l a r c o n c l u s i o n s a r e r e a c h e d i n W i l l i a m s ( 1 9 7 1 ) .)
I n c i d e n t a l l y , n o t e t h e good, s h e t a l k e d a b o u t u s u a l l y q u i t t i n g h e r j o b a f t e r a b o u t o n l y t h r e e weeks.
CHAPTER IV
REMARKS ON OBJECT DELETION
1.
"Object deletion" and the internal structure of purpose
clauses.
There are a number of reasons for analyzing purpose
infinitives as sentential complements.
For examples, we
have seen that the entire infinitive phrase can serve as the
postcopular focus constituent in pseudo-cleft and cleft
sentences:
1.
what he bought the piano for was for Jane to
practice on
it was for Jane to practice on that he bought the
piano
Furthermore, the infinitive phrase,
behaves as a semantic unit.
NP,
including the for-
Consider examples like the
following:
2. a. he bought the piano more for Jane to practice on
than for anything else
b.*these problems were more difficult for the teachers to invent than for the students to solve
c. he bought the piano especially for Jane to practice on
2. d.
this prdblem was especially difficult for Bill
to solve
In (2a), the terms of comparison or contrast are for Jane
to practice on and -for anything else.
~
because for Jane to practice on
semantic unit.
This is possible
is interpretable as a
Note that this contrasts with (2b):
In
(2b), for the teachers to invent and for the students to
,
solve are not permissable terms of comparison, because
they are not semantic units.
It has been argued by several
linguists (including Joan Bresnan and Lasnik and Fiengo)
that the for NP VP sequence in a sentence such as (2b)
-
must be analyzed as a dative for-PP followed by a VP rather
than as a sentential c~mplement. Since, on this analysis,
there is no node which dominates the for NP VP sequence
exclusively, the analysis predicts that it cannot be interpreted as a semantic unit.
(2c, d)
.
This contrast also shows up in
In (2c), especially can modify the whole infini-
tive phrase inclusive of for Jane; otherwise put, the adverb
especially can modify the entire purpose clause, for Jane to
practice on.
By contrast, in (2d), especially can modify
either for Bill-70rto solve, but not the sequence including
both:
3.
this problem was especially difficult for Bill to
solve
--
but not so much for the rest of the class
3.
this problem was especially difficult for Bill
to solve, though he had very little trouble formulating it
*this problem was especially difficult for Bill to
solve, though not so much for his teacher to formulate
It appears that an adverb like especially, exclusively,
principally, particularly, etc., can modify only semantic
units.
If we analyze the for NP VP sequence in the purpose
clause example (2c) as a sentence, it will clearly function
as a semahtic unit, allowing the interpretation available
in ( 2 c )
which is not available in (2d), where the for NP VP
sequence, being analyzed as a PP VP sequence, does not
function as a semantic unit.
We can approach this another way.
Consider the follow-
ing :
4. a. *I built this harpsichord more for Alice to practice on than for Bill
b.
these problems were more diffi~ul~t
for the
teachers to solve than for the students
c. *I built this harpsichord especially for Alice
to practice on and not so much for Bill
d.
these problems were especially difficult for
the teachers to solve, but not so much for the
173
4. d.
students
(I use upper case letters to indicate contrastive stress.)
In (4a) and (4c), the sequence for Alice cannot function as
a term of contrast, whereas in (4b) and (4d), the sequence for the teachers can.
clearcut.)
(The judgments are not very
This contrast is immediately accounted for if
for c lice is not a constituent while for the teachers is
-
analyzed
as a PP.
For Alice could be either a COMP NP
sequence or Alice could be the subject of the sentential
object for.
(See below.)
Then, only for the teachers
is interpreted as a semantic unit and hence is a permissable
term of contrast.
The oddity of (4a, c) is analagous
to the oddity of an:example
such as
5. *I wanted more for Bill to come than for Harry
*I especially hoped for Bill to come, though not
so much for Harry
where for Bill is a COMP NP sequence rather than a for-PP.
Another fact which testifies to the sentence hood of
purpose clauses is that the subject of the purpose clause
is one of the possible syntactic positions of the Themecontrolled NP, as in
6.
I borrowed Fidoi
the afternoon
-i to watch over my children for
Given that purpose clauses are sentential complements,
there would appear to be an immediate problem for any analysis that attempts to account for the anaphoric relationship
in a sentence like (7) in terms of a rule which relates the
two NP's by deletion or interpretation; i.e., a.rule which
crucially involves the controlling NP and the variable NP
inside the purpose clause.
7. I bought the cari for Sam to drive around in
-
Any such rule would be in clear violation of Chomsky's
(1973) Specified Subject Condition, since Sam is the specified syntactic subject of the purpose clause.
The Speci-
fied Subject Condition (SSC) is stated in (8):
8.
No rule can involve X I Y in the structure
. . . x .. .
[
. ..
z
. ..
-wuv . . .
1
..
*
OQ'
where Z is the specified subject of WYV
Noam Chornsky (personal communication) has suggested to
me that there would be no problem here if the structure of
-
the for-phrase in (7) were analyzed as in (9).
9.
I bought the cari:[pp for [
Sam to drive around
'red
'red is what Chomsky calls "reduced sentence", which figures
in the base rules as follows:
S .-jCOMP
'red 4 N P
AUX VP
Given thatocin (8) is a cyclic category and that Chomsky's
analysis assumes that
S,
rather than Sred is the domain of
cyclic rules, it follows that the Specified Subject Condition does not apply in the case of (7) assuming (9) to be
the correct structural analysis .'I
*
Following Chomsky's suggestion, we hypothesize that PP
can have the expansion indicated in (11).
Rule (11) provides the internal structure of purpose clauses
and, presumably, of the for-phrase complements of adjectives
like ready.3
The rest of this chapter will be devoted to remarks on
other constructions that manifest complement object deletion.
We will first consider tough-predicates and then move
on to an analysis of too and enough constructions.
2,
Remarks on tough-class predicates.
Interest in the tough-predicates began with the problem
of accounting for the difference between:
1.
John is easy to please
2.
John is eager to please
John is the understood object of please in (1) but the understood subject of please in (2).
The traditional generative
analysis posits a movement rule which derives (1) from,
-
it is easy to please John by promoting the ~ b j e c tJdhn to
the position of matrix subject.
(henceforth L
&
Lasnik-and Fiengo (1973)
F), who challenge the traditional movement
analysis for (I), account for the difference as follows:
(see their paper fox a summary of the issues): (1) There is
a rule of Object Deletion (OD) which deletes the objects
of infinitive complements under identity with some NP in
the matrix sentence.
Subject Condition
--
(2) OD is subject t o ~ t h eSpecified
more specifically, the 'strong' form of
the condition, according to which the complement subject is
specified unless it is controlled by the NP containing the
term X in (8): (3) OD is obligatory in VP complements and
optional in S complements.
-
( 4 ) Predicates like -easy, hard,
etc. take VP complements, a fact of subcategorization.
(5)
Eager takes a sentential complement, and OD, which is
optional in this case, is prohibited from applying by the SSC.
Consider the following examples:
3.
John is easy for Max to live with
4. *John is eager for Max to live with
John is eager for there to be books about him
5.
6. *John is easy for there to be books about him
7 , *John is eager for there to be books about
8. *John is easy for there to be books about
-
(3) is grammatical; the complement of easy is interpreted as
a PP VP sequence, and O D applies obligatorily into the VP
complement.
(4) is excluded, since eager takes a sentential
complement, and the application of OD is prohibited by the
SSC.
(5) is perfectly good, with eager taking a sentential
complement with no application of OD.
(6) is excluded, since
There-Insertion
requires a sentential anthe application of alysis of the comDlement, while easy is subcategorized for
a (PP) VP complement.
(41,
(7) is out for the same season as
and (8) is out both because the subcategorization of
easy does not allow for a sentential complement and because
the rule of OD has applied over a specified subject.
for-phrases developed in this theGiven the analysis of sis, the deviance of (4) and ( 7 ) has an alternative description.
Eager, like other psychological predicates, takes a
i
for-phrase
-
complement.
It might therefore seem possible for
(4) to be generated with the structure shown in ( 9 ) .
9.
John is eager [ p p for [
Max to live with PRO] ]
'red
17 8
Given such a s t r u c t u r e , t h e OD r u l e would n o t be blocked by
t h e SSC.
But r e c a l l t h a t we argued i n Chapter I1 t h a t t h e
a p p l i c a b i l i t y of t h e r u l e e f f e c t i n g OD i s s u b j e c t t o condit i o n s on t h e s e m a n t i c r e l a t i o n between t h e c o n t r o l l i n g N P
-
and t h e f o r - p h r a s e .
predicates,
I n t h e d i s c u s s i o n of r e a d y and r e l a t e d
it was o b s e r v e d t h a t p s y c h o l o g i c a l p r e d i c a t e s
do n o t mark a s e m a n t i c r e l a t i o n between t h e i r s u b j e c t s and
for-phrase
of OD.
complements t h a t c o r r e l a t e s w i t h t h e p o s s i b i l i t y
T h e r e f o r e , t h e a p p l i c a t i o n of t h e r u l e i n ( 9 ) i s
blocked by s e m a n t i c c o n d i t i o n s , and ( 4 ) i s o u t .
Clearly,
we c a n n o t a s s i g n ( 3 ) a s t r u c t u r e a n a l o g o u s t o ( 9 ) .
p r e d i c a t e s do n o t t a k e f o r - p h r a s e complements.
Tough-
4
There i s a good d e a l of p l a u s i b i l i t y t o t h e c l a i m t h a t
t h e i n f i n i t i v e complements t o t o u g h - p r e d i c a t e s t h a t have
m i s s i n g o b j e c t s a r e VP's r a t h e r t h a n s e n t e n c e s .
For i n s t a n c e ,
it h a s been shown i n Bresnan (1971) t h a t t h e f o l l o w i n g k i n d
of s e n t e n c e i s ambiguous:
it would be good f o r John t o p l a y w i t h S a l l y
10.
( 1 0 ) can be b r a c k e t e d i n two ways, a s f o l l o w s :
11. a .
b.
it would be good [ f o r John] [ t o p l a y w i t h S a l l y ]
it would be good [ f o r John t o p l a y w i t h S a l l y ]
I n ( l l a ) , f o r John i s a PP f u n c t i o n i n g a s a d a t i v e complement
-
t o good,
w h i l e i n ( l l b ) , f o r John t o p l a y w i t h S a l l y i& a
179
sentential complement.
The two possible readings of (10) are
clearly distinguished in the two possible unextraposed versions of that sentence:
12. a.
to play with Sally would be good for John
b.
for John to play with Sally would be good
As is well known, good is one of the tough-predicates which
allows a-missing complement object, as in (13).
13.
Sally would be good for John to play with -
But, in this kind of example, the only possible bracketing
is of the-type where for John is a PP:
14. a.
Sally would be good [for John] [to play with-1
b. *Sally would be good [for John to play with-]
As evidence for this, consider the following:
When good
takes a sentential complement, as in (llb), it can also
take a for-PP dative complement, as in (15).
15.
it would be good - for the family - for John
to play with Sally
Notice, however, that there is no grammatical "missing object"
version of (15):
16.
*Sally would be good for the family for John to
play with ,
for-phrase for the family forces us to
The presence of the construe John as the syntactic subject of the infinitfve
complement; and, under these circumstances, the hypothetical
object-deletion or object-promotion rule is prohibited from
applying.
'.&hat,.
the object of the for-PP dative
Further ,.-mote
complement of good is restricted to nouns that can be
animately conceiized.
17.
Thus, in a sentence like
it would be good for the chalk to stick to the
blackboard
we would normally coqstrue the chalk as the syntactic subject of the infinitive phrase:
18. a.
for the chalk to stick to the blackboard would
be good
b. *to stick to the blackboard would be good for the
chalk
(Compare (18b) withfGtostick to the point would be good for
John.)
Observe now that (19) is odd in exactly the same way
as (18b).
19.
*the blackboard would be good for the chalk to
stick to -
This fact shows that, as in (18b), for the chalk is analyzed
for-FP complement in (19).
as a dative -
The oddness of
these examples is due to the fact that the chalk does'not
satisfy the selectional restrictions on the object of the
dative for-PP.
It would be plausible at this point to follow Bresnan
and suggest that the transformation operates only into VP
complements and, therefore, that the infinitive phrase in
(13) is a VP rather than-an S :
20. Sally would be good ipp for John1 LVp
to play with
-1
However, there is clearly at this point an alternative.
No-
tice that our evidence really shows only that when the object of the infinitive complement of a tough-predicate is
for NP sequence appearing after the adjective
missing, a -
-
head must be construed as a dative 'for-PP complement rather
than as a complementizer followed by the syntactic subject
of the infinitive.
One could propose that the infinitive
phrase to play with in (13) is a~sehtencewith a syntactically empty subject (interpreted as coreferential with the
object of the dative for-PP) and formulate the objeet-deletion or object-promotion rule so that it will block if the
subject of the infinitive phrase is filled.
we could formulate OD roughly as follows:
To illustrate,
21.
NP
Pred
1
2
(PP) V*
3
(P) NP
4
5
6 + 1 2 3 4 5 $
In (21), V*-represents an arbitrarily long string of verbs
(see Bresnan (19713266, 276) )
.
Note that (14b) and (16)
do not satisfy the structural description of the Yule (21)
and are consequently excluded.
Suppose we analyze (13) as
follows:
Sally would be good for John
[S
for PRO
[VP to play with NP] 1
Nohice that (22) could satisfy the structural description of
(211, if we assume that Equi deletes7COMP NP and that it is
ordered before OD.
However, assuming a rule like OD or
Bresnan's Object Shift (an object-promotion alternative),
the application of (21) in (22) would still be blocked by
the strong form of the SSC, because Sally, the antecedent
of the object of with, does not control the complement subject.
(Incidentally, the assumption that Equi would apply
before OD is natural if the complement structure of good is
dominated by the AP node and the AP node is cyclic.)
There are certain facts which may indicate that a
promotion rule rather than>a deletion rule is involved,
contrary to L
&
F's claims.
contrast the examples
On p. 24 of their paper, L
&
F
2 3 . a . *prime numbers a r e e a s y t o prove E u c l i d ' s
theorems a b o u t
b.
prime numbers a r e e a s y t o prove theorems
about
The c o n t r a s t between t h e s e examples i s supposed t o i l l u s t r a t e
t h e workings of t h e S p e c i f i e d S u b j e c t C o n d i t i o n :
OD i s
blocked i n ( 2 3 a ) b e c a u s e t h e N P - c o n t a i n i n g t h e o b j e c t t o be
d e l e t e d has a s p e c i f i e d s u b j e c t , Euclid.
misses t h e point.
But c l a a r l y t h i s
Observe t h a t i n ( 2 3 b ) , p r o v e theorems
must be a n a l y z e d a s a p h r a s e , e q u i v a l e n t t o t h e o r i z e .
prime numbers a r e e a s y to [prove theorems]
24.
about
prime numbers a r e e a s y t o t h e o r i z e a b o u t
(23b) c a n n o t p a r a p h r a s e , theorems a b o u t prime numbers a r e eas y t o prove; i . e . ,
it c a n n o t have t h e s t r u c t u r e i n d i c a t e d i n
(25).
*prime numbers a r e e a s y t o prove [theorems
25.
about]
Note t h a t i f w e s u b s t i t u t e f o r p r o v e a v e r b t h a t c a n n o t form
a p h r a s e w i t h theorems, t h e s e n t e n c e i s ungrammatical.
2 6 . a . *prime numbers a r e e a s y t o comprehend theorems
about
b.
theorems a b o u t prime numbers a r e e a s y t o comprehend
184
Consider further the differences between the a- and bexamples of (27) and (28)..
27. a.
Bill is hard to take pictures of
b. *Bill is hard to look at pictures of
28. a.
this topic would be difficult to write a book
about
b. *this topic would be difficult to read a book
about
Take pictures and write a book are taken as phrases in (27a)
and (28a) respectively, while look at pictures and readba
book in (27b) and (28b) cannot be taken as phrases.
29. a.
Bill is hard fo.>[take pictures] of
b. *Bill is hard to look at [pictures of]
30. a.
this topic would be difficult,to [write a
book] about
b. *this topic would be difficult to read [a book
about]
The data indicate that the rule involved here is subject
to the A-over-A condition.
We would consequently not need
to appeal to the SSC to account fox the ungrammaticality of
(23a). This observation would seem to favor a movement
rule over a deletion rule.
For examfple, if, corresponding
to (29b), we had the underlying form,
31.
Billi be hard [Vp to look at [Np pictures
assuming a deletion analysis, the most inclusive NP inside the VP complement to which the rule could apply is
the object of pictures of.
Thus, the A-over-A Condition
would not block the application of the deletion rule,
predicting that (29b) should be grammatical.
On the other hand, on a movement Alternative, the
.
underlying form corresponding to (29b) would be (32)
be hard
32.
LVp
to look at [Np pictures of
The most inclusive NP to which the movement rule can apply
in (32) would be pictures of Bill.
Therefore, from (32),
only (33) would be generated, and (29b) would be excluded.
33.
pictures of Bill are hard to look at
The evaluation of this argument is complicated by the
fact that the relevant observations are related to facts
about the distribution of indefinites like someone, something.
For example, compare the a- and b-examples of (34)
and (35).
I
186
34. a.
he was working through (Euclid's) proofs about
trianales
b. *he was working through (Euclid's) proofs about
something
35. a.
he was proving theorems about triangles
h.
he was proving theorems about something
The examples indicate that when the about-phrase forms a
phrase with proofs, the object of about cannot be unNotice that in (35b),
stressed indefinite something.
proving theorems must be a phrase.
Compare, *he was memor-
izing theorems about something.
A proponent of the deletion alternative for tough-predicates could argue that the delet@d object is PRO in
underlying structure and that the distribution of P R O is tied
to the distribution of indefinites like someone, something.
In this case, (31) would be revised to (36).
36.
Bill be hard LVp
PRO]
to look at [Np pictures of
I
The deletion rule would make P R O coreferential with the subject of hard and delete it.
The ungramaticality of (29b)
on this alternative would be accounted for in terms of constraints on the distribution of P R O related to constraints
on the distribution of indefinites like someone, ~ o m e t h i n g . ~
(See Oehrle (1974) for relevant discussion.)
3-
Too and
-
enough constructions.
At this point, I would like to turn our attention to
too
and enough constructions, with the aim of attempting to
determine whether or not the infinitive complements to
these degree modifiers are sentential.
Notice that too and
for-phrase complements.
enough are subcategorized for -
1. this music is too slow for modern dancing
Nixon is too right-wing for my vote
Fifivwas not good enough for first prize
he's too short for a Watusi
the dean
considered her- intelligent enough for a
full scholarship
This raises the possibility that the infinitive complements
for-phrases with reduced sento these detree modifiers are tential objects.
(See rule (11) of section one.)
Let's consider the nature of the rule effecting object
"deletion".
Consider the following examples.
2. the statuei was too small -i
to attract attention
3. the statuei was too small for anyone to notice
-
It has been proposed that the deletion rule operating in (2)
is distinct from the one operating in ( 3 ) :
for (22, there
188
would be a rule of Equi-NP deletion deleting the subject of
the infinitive complement under identity with the subject of
the matrixxsentence, while for ( 3 ) , the object of notice
would be deleted by a rule d f OD, as in the analysis outlined by L
&
F.
The distinctness of the two rules is ob-
servable from the fact that OD operates only when the degree modifier modifies a predicative, while the Equi rule is
indifferent to whether the degree modifier modifies a predicative, an adverb, or a bare Q like Much. (Cf. Bresnas
1972).
4.
the statuei was too obviously obscene -p:to attract attention
5. *the statuei was too obviously obscene for anyone
to take notice of 6.
Maryi runs too -fast
to see what's happening
around her
7. *Maryi runs too fast for me to keep up with 8.
Homeri eats too much
--to lose any weight
9. *Homeri eats too much for Jim to keep up with Notice that for (5), (7), and (9), if we have full pronouns
instead of the deletion sites, the sentences are grammatical.
10.
the statuei was too obviously obscene for anyone
to take notice of iti
11.
Maryi runs too fast for me to keep up with heri
12.
Homeri eats too much for Jim to keep up with himi
,
.r
Also, with regard to ( 7 ) , if fast is an adjective, OD is
possible:
13.
-
when it comes to running, Maryi is too fast for
me to keep up with
While these observations are correct as far as they go, the
facts are actually slightly more complicated.
The problem
is that, in the examples considered above, the subject of the
AP is the subject of the matrix sentence.
In sentences
where the subject of the AP is a constituent of the VP, additional relevant observations come to light.
14.
Mary made the statuei too small
-
to attract
-
to attract
attention
15.
Mary
made the statue too small
i
attention
16.
Mary made the statuei too small for anyone to
take notice of -
17. *Maryi made the statue too small for anyone to take
notice of
18.
Maryi made the statue too small for anyone to notice heri
190
From (14) and (15), we see that-there are two possibilities
for the deletion of the complement subject:
It can be con-
trolled either by the subject of the,AP or by the subject of
the matrix sentence.
Further, if OD applies, neither the
subject of the AP nor the subject of the matrix sentence can
control the complement subject.
19.
In the sentence,
Mary made the statuei too small to notice
the subject of the infinitive is unspecified.
-
If OD does not
apply, the subject of the infinitive can be controlled by
the matrix subject, as in
20.
Maryi made the statue too small
j
-
to notice it
j
These examples point up a difference between too and
enough constructions and tough-predicate constructions:
only in the former is it possible for the subject o f t h e AP
to control the complement subject.
Corresponding to (2)
and (14), we do not find any sentences like
21. *the statuei is hard
- to attract attention
*Mary made the statuei hard
to attract
attention
In accsun4ing for this, we could adopt L
&
F's suggestion
too
that there are two alternative subcategorizations for and enough.
That is, they can take full sentential comple-
191
ments or (PP) ;VP complements.
When the complement is sen-
tential, we have ~complenlentsubject deletion, as in (2),
(4),
(6),
( a ) , (14), and (20). When the complement has a
(PP) VP structure, we get complement object deletion (OD).
For tough-predicate constructions, only the latter type of
subcategorization is available.
Hence, only OD is possible.
Now notice that there seems to be a generalization concerning the deletion of the complement subject and the detoo and enough construcletion of the complement object in tions.
too and
We saw above that OD could not apply if -
enough did not modify a predicative.
Thus, (22a) is im-
possible.
22. a. *Mary made the statuei too slowly for anyone tonotice
b.
-
Mary made the statuei too slowly for anyone to notice iti
But it is significant that the statue is also prohibited
too
from controlling the complement subject as well, when does not modify a predisative.
Thus, compare (14) and (15)
with
23.
*Mary made the statuei too slowly
tract attention
to at-
192
24.
Mary i made the statue too slowly
to attract
-a
attention
Quite clearly, then, the subject of the AP can control
either the subject or the object of the complement of too
and enough.
That is, when the infinitive complement is
part of a predicative expression, the complement subject or
object is controlled by the NP which is the subject of the
predicative expression.
(22a) and (23) do not work because
Adverbial Phrases are not predicative.
(24) is grammatical
because there is an Equi rule which assigns control of the
complement subject to the matrix subject.
This Equi rule
operates when the infinitive complement is not part of the
.
predicative expression, as in (15) and (20)
It does not
operate in (19) because the infinitive phrase of that example is part of the predicative expression.6
In the light of these facts, it would seem plausible
to consider-"analternative to L
&
F's proposal.
The gen-
eralization noted between complement subject deletion and
complement object deletion can be accounted for if it is
assumed that when the infinitive complement is part of the
predicative expression, it has the internal structure given
by rule (11) of the first section.
On this analysis, the
structures of (14) and (16) would be as given in (25) ahd
(26), respectively.
25. Mary made the statuei too small [ p p for [
'red
-
i
to attract attention]]
26. Mary made the statuei too small [pp for [
any'red
one to take notice of - 1 1
An obligatory rule of pronominalization makes the subject
of the predicative expression coreferential with the subject of the complement in (25), and with the object of the
complement in (26). Given such an analysis, the application of this rule in (26) would not violate the SSC.
The two major arguments that L
&
F adduce against a
sentential analysis for such complements involve the fact
that two transformations, Passive and
here-Insertion,
which operate only on sentences, are prohibited from applying to the complements of too and enough if OD applies.
Thus, the following examples are ungrammatical:
27. *John is not famous enough for there to be a book
about
-
28. *John is too disoriented for the parade to be led
by
-
It is not at all clear, however, that the unacceptability
of such examples is due to syntactic violations.
For ex-
ample, as Roger Higgins pointed out to me, examples like the
first sentence of (27) must be exc3rrded even if There--.-!Insertion does not apply:
194
28.
*John is not famous enough for a book to be about
Or consider pairs like the following:
29. a. *$his garden is not big enough for there to be
people in
b. *this garden is not big enough for people to be
in
(29b) is out, even though for people qualifies as a dative
for-phrase, since
-
its object is animate.
It seems, in fact, quite likely that sentences like - those in (27)-(29) are to be excluded on semantic grounds.
Purpose clauses in many cases manifest the same resistance
to There-Insertion:
"
pm
30. a.
I wrote this play for Kazin to review
b. *I wrote this play for there to be reviews of
31. a.
I bought this cottage for my guests to stay in
b. *I bought this cottage for there to be guests in
Note, however, that certain verbs like pick and choose do
allow There-Insertion complements:
32.
I chose this play for there to be review of
I picked this topic for there to be debates
about
-,
195
Much the same situation holds for Passive.
Sentences
like the following are not acceptable:
"Socrates is intelligent enough for us to be
33.
convinced by
*the police are too stupid for Bonnie to be captured by
*this music is too cacaphonous for me to be put
to sleep by
However, it is possible to find passive examples of OD corn?
plements to too and enough.
this word is too short for the stress to be
34.
,'
place on
this stuff is not nutritious enough for a baby
to be weaned on
Such examples as in (34) seem quite acceptable. 7
Concerning purpose clauses, there are many cases where
a passive infinitive complement is not acceptable, depending on the matrix verb, the content of the infinitive phrase
etc.
While (35b) is out, (36b) seems perfect19 )acceptable:
35. a.
we needed Billi
-i
to lead the parade
b. *we needed Billi for the parade to be led
by -i
,
36. a.
b.
w e chose B i l l i -i
we c h o s e B i l l i
A s we saw w i t h
t o lead t h e parade
f o r t h e p a r a d e t o be l e d by
-
here-Insertion, choose imposes fewer r e s t r i c -
t i o n s on i t s p u r p o s e c l a u s e complements t h a n o t h e r m a t r i x
verbs.
Verbs t h a t t a k e purpose c l a u s e complements impose
d i f f e r e n t r e s t r i c t i o n s on t h e f u n c t i o n a l r o l e t h a t t h e purpose c l a u s e can c h a r a c t e r i z e .
For i n s t a n c e , t h e o b j e c t of
u s e c a n n o t c o n t r o l t h e o b j e c t of t h e complement, b u t t h e
o b j e c t of
buy can :
a.
I used t h e s e g l a s s e s i t o r e a d w i t h
b.
I bought t h e s e g l a s s e s
b.
I bought t h i s n o v e l i
-
t o read with
i
-i
3 8 . a . *I used t h i s n o v e l i t o r e a d - on t h e t r a i n
I t seems t o m e t h a t
t o reqd
-i
on t h e t r a i n
( 3 8 a ) i s u n a c c e p t a b l e simply because i t
i s n o t p r o p e r t o speak of a n o v e l a s b e i n g used i n t h e a c t
of r e a d i n g when it i s b e i n g r e a d .
On t h e o t h e r hand, when
one r e a d s w i t h a p a i r of g l a s s e s , it i s p e r f e c t l y p r o p e r t o
speak of u s i n g t h e g l a s s e s i n t h e a c t of r e a d i n g .
R e t u r n i n g f o r a moment t o T h e r e - I n s e r t i o n ,
that
notice
here-Insertion complements a r e c o m p a t i b l e w i t h e a g e r
(the predicate t h a t L
& F
but o f t e n not with ready.
c o n t r a s t with t h e tough-predicates)
Compare:
197
39. a.
I am eager for there to be books about me
b. ?I am ready for there to be books about me
40. a.
I am eager for there to be guests in my home
b. *I am ready for there to be guests in my home
The contrasts in (39) and (4)) do not involve object deletion at all.
Passive complements contrast with ready and
eager in a similar way:
41. a.
I am eager for John to be arrested by the police
b. ?I am ready for John to be arrested by the police
The following, however, seem equally good:
42. a.
I am eager for my film bo be shown on television
b.
I am ready for my film to be shown on television
Thus it seems clear that There-Insertion and Passive can
interact in (sometimes subtle) ways with the interpretation
of the complement to diminish acceptability.
It seems
likely that the violations represented in (27) and (33)
are of this type.
The fact that judgments in this area
are often so delicate and change with the lexical content
of the infinitive complement suggests that the violations
are not of a general structural sort, as L
&
F's OD anal-
ysis claims.
We are led now to consider the question whether there
are any arguments in favor of analyzing OD complements of
198
t o o and
the for
enough a s s e n t e n t i a l .
Certain f a c t s indicate t h a t
NP sequence of t h e OD complements of t o o and enough
for-PP's.
do n o t behave a s d a t i v e -
For example i n tough-
predicate constructions, the for-PP i s s e m a n t i c a l l y a
d a t i v e complement and i t s o b j e c t N P i s l i m i t e d t o N P ' s which
can be a n i m a t e l y c o n c e i v e d .
-
However, t h e r e i s no such r e -
s t r i c t i o n i n t h e c a s e of t o o and enough.
Thus, c o n s i d e r
t h e f o l l o w i n g examples:
43.
a.
t h e s u r f a c e was n o t pDrous enough f o r t h e c h a l k
t o adhere t o
b.
t h e s a l t c r y s t a l s were t o o l a r g e f o r t h e w a t e r
t o b r e a k down
c.
t h e l i q u i d was t o o v i s c o u s f o r t h e sponge t o
absorb
4 4 . a . * t h e s u r f a c e would n o t be e a s y f o r t h e c h a l k
t o adhere
to
b . * t h e s a l b c r y s t a l s would be u s e f u l f o r t h e w a t e r
t o breqk down
c . * t h e l i q u i d would be dangerous f o r t h e sponge t o
absorb
N a t u r a l l y , f o r t h o s e t o u g h - p r e d i c a t e s which a r e a l s o s u b c a t e g o r i z e d f o r s e n t e n t i a l i n f i n i t i v e complements, t h e r e i s
no animacy r e s t r i c t i o n on t h e complement s u b j e c t N P :
45. a .
it would be u s e f u l f o r t h e w a t e r t o b r e a k
down t h e s e c r y s t a l s
b.
it would be dangerous f o r t h e sponge t o a b s o r b
this liquid
F u r t h e r m o r e , i f t h e u n d e r l i n e d N P ' s i n ( 4 4 ) a r e r e p l a c e d by
a n i m a t e s , t h e s e n t e n c e s analogous t o ( 4 4 ) become gramrnat i d a l , and t h e ones a n a l o g o u s t o ( 4 5 ) become ambiguous:
46. a.
t h i s s u r f a c e would n o t be e a s y f o r a f l y t o
stick t o
b.
t h e s e s u b s t a n c e s would be u s e f u l f o r t h e l i v e r
t o b r e a k down
c.
t h i s c h e m i c a l would be dangerous f o r t h e c e l l s
t o absorb
4 7 . a.
i t would be u s e f u l f o r t h e ' l $ v & r ~ t o ' b r e a kdown
these substances
1. t o b r e a k down t h e s e s u b s t a n c e s would be use-
f u l for the liver
2 . f o r t h e l i v e r t o b ~ e a kdown t h e s e s u b s t a n c e s
would be u s e f u l
b.
it would be dangerous f o r t h e c e l l s t o a b s o r b
t h i s chemical
1. t o a b s o r b t h i s c h e m i c a l would be dangerous
for the c e l l s
2.
f o r t h e c e l l s t o a b s o r b t h i s chemical would
be dangerous
200
Second, as we observed before, contrary to what L
&
F
suggest, it is possible to find some examples of passive contoo and enough
structions in OD complements of -
where,
moreover, the sentences are synonymous with their active
counterparts.
48. a.
this product is not nutritious enough for a baby
to be weaned on
b.
this product is not nutritious enough to wean
a baby on
49. a.
this word is too short for the stress to be
placed ;on
b.
this word is too short to place the stress on
Compare these examples with ones like the following.
50. a.
it could be dangerous to wean a baby on this
product
b.
it could be dangerous for a baby to be weaned
on this product
The sentences of(5B) are synonymous only when the complement
of (50b) is construed as a sentence.
If for a baby is
analyzed as a PP complement, there is a direct implication
of danger to the baby, which is lacking in (50a).
Thus, the
a
facts indicate that active/passive synonymy holds only if baby is taken to be the syntactic subject of be weaned on.
201
S i n c e t h i s i s n o t p o s s i b l e i n t h e OD a n a l o g t o ( 5 0 b ) , we
-
would p r e d i c t t h a t (51b) i s n o t synonymous w i t h ( 5 1 a ) , t h e
OD a n a l o g t o
51. a .
b.
(50a).
This prediction i s , i n f a c t , correct.
t h i s p r o d u c t c o u l d be dangerous t o wean a baby
t h i s p r o d u c t c o u l d be dangerous f o r a baby t o
be weaned on
-
S i n c e i n ( 5 1 b ) , f o r a baby must be c o n s t r u e d a s a d a t i v e complement, we i n f e r t h a t t h e r e i s danger f o r t h e baby i t s e l f .
There i s no such i n f e r e n c e from ( 5 1 a ) ,
From t h e s e c o n s i d e r -
a t i o n s , we c o n c l u d e t h a t t h e a c t i v e / p a s s i v e synonymy of t h e
examples i n ( 4 8 ) and ( 4 9 ) i n d i c a t e s t h a t t h e f o r NP sequences
a r e n o t d a t i v e complements.
One f u r t h e r argument h a s t o do w i t h t h e d i s t r i b u t i o n of
u n s t r e s s e d i n d e f i n i t e someone, anyone.
Some of t h e tough-
p r e d i c a t e s a r e a f f e c t i v e , i n K l i m a ' s s e n s e (Klima ( 1 9 6 4 ) ) ;
any i n t h e i r i n f i n i t i v e complethey allow f o r unstressed -
52.
i t would be h a r d t o f o r c e anyone t o come
i t would be h a r d t o f o r c e someone t o come
it would be dangerous t o f o r c e anyone t o come
i t would be dangerous t o f o r c e someone t o come
The s e n t e n c e s of
( 5 2 ) and t h e s e n t e n c e s of
(53) a r e equally
good, b u t d i f f e r e n * i n meaning.
These p r e d i c a t e s do n o t a l l o w f o r t h e p r e s e n c e of anyone
a s t h e o b j e c t of t h e i r d a t i v e for-PP
54. a .
complements.
it would be h a r d f o r someone t o f o r c e B i l l
t o come
b. * i t would be hard f o r anyone t o f o r c e B i l l t o
come
55. a .
i t would be dangerous f o r someone t o f o r c e
B i l l t o come
b.
i t would be dangerous f o r anyone t o f o r c e B i l l
t o come
S i n c e h a r d d o e s n o t t a k e a s e n t e n t i a l complement,
ungrammatical and ( 5 4 a ) i s unambiguous.
(54b) i s
On t h e o t h e r hand,
(55b) i s a c c e p t a b l e i f anyone i s t a k e n a s t h e s y n t a c t i c subj e c t of t h e i n f i n i t i v e p h r a s e , s i n c e dangerous can t a k e
s e n t e n t i a l camplements.
56. a . * f o r anyone t o f o r c e B i l l t o come would be hard
* t o f o r c e B i l l t o come would be h a r d f o r anyone
b.
f o r anyone t o f o r c e B i l l t o come would be dangerous
* t o f o r c e B i l l t o come would be dangerous f o r
anyone
( 5 5 a ) seems t o m e t o be ambiguous; someone can be c o n s t r u e d
203
as the object of the preposition for, or as the syntactic
subject of the infinitive phrase.
The distributional pat-
tern for unstressed someone, anyone is shown in the following examples.
57. a.
it would be .[haPd for someone] to adhere to
such a diet consistently
b. +it would be [hard for anyone] to adhere to
such a diet consistently
58. a.
it would be [dangerous for someone] to contract
this disease
b. it would be [dangerous for anyone] to contract
this disease*.
c.
it would be dangerous [for someone to contract
this disease]
d.
it would be dangerous [for anyone to contract
this disease]
59. a.
this diet would be [hard for someone] to adhere to consistently
b. *this diet would be [hard for anyone] to adhere to consistently
60. a.
this disease would be [dangerous for someone]
to contract
b. *this disease would be [dangerous for anyone]
to contract
In conclusion, the relevant observation is that the a£fective tough-prediaates allow for the occurrence of unstressed anyone only within the infinitive complement.
In
particular, anyone in the above examples must be analyzed
as the syntactic subject of the sentential infinitive complefor-PP.
ment and not as the object of a dative Now, sbserve that too (but not --*.
enough) is also affective:
61.
a.
he's too cheap to carry some/any money
b.
he's smart enough-to carry somej*any money
for in tooNotice that unstressed anyone can occur?,after constructions:
62.
a.
the statue was too small for anyone to notice
b.
the problem was too intricate for anyone to
solve
c. *the problem was simple enough for anyone to
,
solve
On the basis of the distributional facts that we have observed
above, the data of (62) argue that anyone is the syntactic
too.
subject of the infinitive complement of -
Notice that
if we postpose the for NP in such examples, forcing a PP
analysis, the resulting sentences are ungrammatical.
(I do
not mean to imply that khere is any postposing transformation
a t work h e r e . )
63.
a . * t h e s t a t u e was t o o s m a l l t o n o t i c e , f o r anyone
b. * t h e problem was t o o i n t r i c a t e t o s o l v e , f o r
anyone
To make t h e examples of
( 6 3 ) a c c e p t a b l e , we would have t o add
a n o t h e r a f f e c t i v e element t o a l l o w f o r anyone a s t h e o b j e c t
-
of f o r :
64.
a.
t h e s t a t u e w a s n ' t t o o s m a l l t o n o t i c e , , f o r anyone
b.
t h e problem w a s n ' t t o o i n t r i c a t e t o s o l v e , f o r
anyone
c.
it was s u r p r i s i n g t h a t t h e problem was t o o
i n t r i c a t e t o s o l v e , f o r anyone
I n c o n c l u s i o n , t h e d a t a examined above seem t o i n d i c a t e t h a t
t o o can
-
take a (reduced) s e n t e n t i a l f o r - p h r a s e complement
when t h e complement h a s
a missing o b j e c t .
FOOTNOTES
CHAPTER IV
1.
I had been assuming that the underlying structure of (7)
was
-
i.
I boaght the cari
in
Lpp
for [S Sam to drive around
-i I 1
with the preposition for being ultimately deleted (see
Bresnan (1972)).
The only reason for this assumption was
that a full sentence complement appeared to show up in the
pseudo-cleft sentence, what,I bought the car for was for Sam
to drive around in.
But Chomsky pointed out to me the
acceptability of sueh a pseudo-cleft sentence as, what I
bought the car f o was
~
for pleasure trips,
-- with a forphrase in focus position.
for was pleasure trips.)
(Compare, what I bought the car
Thus, it would seem possible to
analyze the infinitive phrase in focus position as a full
for-phrase
-
:
rather than a simple sentential complement.
Therefore, pseudo-cleft sentences provide no argument against
the structural analysis in (9).
2.
A
Gilbert Harman (personal communication) pointed out a
possible problem with this suggestion.
If a purpose clause
has the structare-indicated in (9), the application of the
-
r u l e of ~ a c h - I n s e r t i o n i n t o p u r p o s e c l a u s e s s h o u l d n o t be
blocked.
But c o n s i d e r :
i. a .
John and B i l l e a c h bought a few o f Sam's books
f o r S a l l y t o read t o t h e other
b.
* J o h n - a n d B i l l bought a few of Sam's books f o r
S a l l y t o read t o each o t h e r
Clearly,
l i t y of
( i b ) c a n n o t be d e r i v e d from ( i a ) .
The ungrammatica-
(ib) i n d i c a t e s t h a t t h e SSC a p p l i e s t o movement i n t o
purpose c l a u s e s and, t h e r e f o r e , t h a t ( 9 ) i s n o t t h e c o r r e c t
s t r u c t u r a l d e s c r i p t i o n of
(7).
T h i s argument i s d u b i o u s , however.
and ( i i b )
Compare ( i i a )
.
ii. a .
John a n d . B i l l e a c h bought a few of Samrs books
t o read t o t h e other
b.
John, alid B i l l bought a few of Sam's books t o
read t o each o t h e r
These two s e n t e n c e s a r e q u i t e d i s t i n c t i n meaning, and it
i s h i g h l y d o u b t f u l t h a t ( i i b ) i s d e r i v e d from ( i i a ) .
If
t h i s i s s o B a c h - I n s e r t i o n must be p r o h i b i t e d from a p p l y i n g
i n t o p u r p o s e c l a u s e s i n d e p e n d e n t l y of t h e SSC.
f o r e , t h e u n g r a m m a t i c a l i t y of
dence a g a i n s t ( 9 )
.
There-
( i b ) does n o t provide e v i -
208
3.
F u r t h e r m o r e , because of o u r o b s e r v a t i o n s a b o u t gerunds
i n Chapter 111, we w i l l r e q u i ~ et h e r u l e s
i.
ii.
PP+P
NP
NP-+VP
t o p r o v i d e t h e i n t e r n a l s t r u c t u r e of g e r u n d i v e purpose forphrases, a s i n
iii.
John bought it Eor.hun$ing s n a r k s w i t h
which would be a n a l y z e d a s
iv.
John bought it [ p p f o r [ NP [VP h u n t i n g s n 3 r k s
1
.
.
with] 1 ]
4.
C o n s i d e r i n c o n t r a s t what I w i l l c a l l e v a l u a t i v e p r e d i d
cates.
E v a l u a t i v e p r e d i c a t e s a r e a c l a s s of p r e d i c a t e s
( i n c l u d i n g some of t h e . t o u g h - p r e d i c a t e s ) which e x h i b i t
" o b j e c t d e l e t i o n " i n t h e i r complements b u t which e n t e r i n t o
c o n s t r u c t i o n s which a r e f o r m a l l y d i s t i n c t from tough-predicate constructions.
Some examples a r e good, a p p r o p r i a t e ,
a c c e p t a b l e , and u s e f u l .
i.
ii.
t h i s g i f t i w o u l d n ' t be a c c e p t a b l e f o r a man t o
g i v e -i t o a woman
such musici o n l y seems good f o r teeny-boppers t o
dance t o
,-
iii.
these toysi are useful for exhausted parents to
keep their children amused with -
Evaluative predicates are subcategorized for for-phrase complements.
iv.
this gift wouldn't be acceptable for a bar mitzvah
v.
vi.
such music only seems good for slow dances
these toys are useful for educational purposes
Observe the similarity in interpretation of these for-phrases
-
and purpose for-phrases.
vii.
viii.
ix.
I purchased this gift for the bar mitzvah
this music was only intended for slow dances
the manufacturer designed these toys for educational purposes
Compare the following with (i-iii):
x.
such thingsi were meant for men to give .i
to women
xi.
I bought this new recordi for all of us to '
dance to -
xii.
we borrowed our neighbor's toysi to keepdzour
children amused with -
210
I t seems, t h e n ,
t h a t t h e i n f i n i t i v e complements i n
(i-iii)a r e i n t e r p r e t e d as purpose c l a u s e s , l i k e t h e comple-
ments i n ( x - x i i ) .
Semantically, evaluative predicates
e v a l u a t e t h e i r s u b j e c t s w i t h r e s p e c t t o t h e purpose o r
f o r c p h r a s e complement:
f u n c t i o n d e p i c t e d by t h e xiii.
a s p i r i n i s good f o r a r t h r i t i s
doctors often prescribe aspirin for a r t h r i t i s
a s p i r i n i s o f t e n used f o r a r t h r i t i s
aspirin is for arthritis
xiv.
such a n i n s t r u m e n t i s n o t a p p r o p r i a t e t o p l a y c l a s s i c a l music on
s h e bought t h i s e x p e n s i v e i n s t r u m e n t t o p l a y
c l a s s i c a l music on
t h i s i n s t r u m e n t h a s n e v e r been used t o p l a y
c l a s s i c a l music on
These c o n s i d e r a t i o n s i n d i c a t e t h a t t h e p r o p e r s t r u c t u r a l ana l y s i s f o r a s e n t e n c e l i k e ( i ) i s r o u g h l y a s g i v e n below.
xv.
t h i s g i f t i w o u l d n ' t be [ADJ a c c e p t a b l e ]
Lpp
for [
'red
a man t o g i v e
- t o a woman] ] ..
R a t e t h a t i n many c a s e s ( a c c e p t a b l e f seems t o be a n except i o n ) , t h e p u r p o s e c l a u s e complement of t h e e v a l u a t i v e p r e d i c a t e c a n be p s e u d o - c l e f t e d :
211
xvi.
what this kind of music is good for is for
teeny-boppers to dance to
what these toys are use=.lulfor is for exhausted
parents to keep their children amused with
Also, as we would expect, there are cases where the subject
of the evaluative predicate controls the complement subject:
xvii.
Johni was not deemed acceptable
-i
to lead
the parade
aspirini is useful i
swelling
to reduce pain and
marijuana i is good -i
to ease nervous tension
In addition, the for-phrase complements of evaluative predicates can be gerundive, as in
xviii.
chantersi are useful for practicing on i
ovensi are good for baking cakes in
this refrigeratori isn't fit for keeping fresh
food in
-
~ h u s ,evaluative predicate constructions provide us
with a further example of "object deletion" into reduced
sentential complements.
The OD rule operates obligator-
ily, as with purpose clauses.
xix. *this gifti wouldn't be appropriate for a guy
to give iti to a girl
*such music i only seems good for teeny-boppers to
dance to iti
*these toysi are useful to keep children amused
with themi
Observe that a sentence like, this music is good for
people to dance to, is ambiguous between a-(tough-predicate
and evaluative-predicate interpretation:
xx.
a.
this music is good
dance to
b.
tpp
-1
this music i s good Ipp
dance to
for people] IVP to
-1 1
for [ S
people to
red
The pseudo-cleft sentence, what this music is good for
is for people to dance :to isolates the evaluative-predicate reading.
5.
Notice that while, *he was looking at Bill's pictures of
someone is impossible, he was looking at pictures of someone
(where pictures of someone is a phrase) is as acceptable as,
he was taking pictures of someone.
It would seem, then, that
the ungrammaticality of sentences like, *Bill is fun to look
at pictures of cannot be accounted for in terms of the distri-
213
b u t i o n of i n d e f i n i t e s .
However, such c a s e s seem t o be
isolated.
I t i s p o s s i b l e t h a t t h e f a c t t h a t t h e s u b j e c t of t h e
6.
i n f i n i t i v e i n ( 1 9 ) i s interpreted a s unspecified w i l l req u i r e t h a t t h e i n f i n i t i v e p h r a s e be analpped a s a V P .
T h a t i s , i t i s p o s s i b l e t h a t t h e s t r u c t u r e of
(19) is,
underlyingly,
i.
Mary made t h e s t a t u e i t o o s m a l l [pp f o r [Vp t o
notice
-i 1 I
rather than
ii.
Mary made t h e s t a t u e i t o o s m a l l [pp f o r '
[s
A
red
t o notice - I 1
with A i n t e r p r e t e d a s unspecified.
What i s a t i s s u e h e r e i s
whether o r n o t t h e r e i s a r u l e t h a t would i n t e r p r e t
b
as
unspecified i n t h e c o n t e x t provided i n ( i i ) .
7.
L
& F
a r g u e t h a t t h e complement of t r y must be a VP
r a t h e r than an
Sf
d
s i n c e , on t h e i r a n a l y s i s , OD i s p r o h i b i t e d
f r m a p p l y i n g i n t o s e n t e n c e s and y e t examples l i k e ( i )
grammatical:
214
i.
the problem was too intricate for John to try
tq solve
-
7
However, it is well known that try can have passive complemerits as in, I tried to be arrested by the police.
Thus,
if try takes a VP complement, Passive is not restricted
to sentences and L
&
F's argument is vitiated.
Otherwise,
try must take sentential complements, and their entire
OD analysis is flawed.
REFERENCES
Bresna0,-'J.1971.
Sentence stress and syntactic trans-
formations.
Language 47.257-281.
Bresnan, J. 1972.
Theory of Complementation in English
syntax. Unpublished doctoral dissertation, M.I.T.
Chomsky, N. 1965.
Aspects of the theory of syntax.
Cam-
bridge, Mass.: M.I.T. Press.
Chomsky, N. 1973.
Conditions on transformations.
A
festschrift for Morris Halle, ed. by S. Anderson
and P. Kiparsky, 232-286.
New York: Holt, Rinehart
and Winston.
Emonds, J. 1970.
Root and structure-preserving trans-
formations.
Unpublished doctoral dissertation,
M.I.T.
Faraci, R.
1971.
A preliminary examination
of for-to
-
purpose clauses in ~nglish. Unpublished paper,
M.I.T.
Faraci, R.
1973.
Use-constructions revisited. Uppub-
lished paper, M.I.T.
Gruber, J.
1965.
Studies in lexical relations.
Unpublished
doctoral dissertation, M.I.T.
Higgins, F.R.
lish.
1973.
The pseudo-cleft construction in Eng-
Unpublished doctoral dissertation, M.I.T.
Jackendoff, R.
1972.
grammar.
Semantic interpretation in generative
Cambridge, Mass.:
Jespersen, 0. 1927.
M.I.T. Press.
A modern English grammar1 on histori-
cal principles.
7 vols.
London: George Allen and
Unwin.
Klima, E.
1964.
Negation in English.
The structure of
language, ed. by J. Fodor and J.Katz, 246-323.
Englewood Cliffs: Prentice-Hall.
Lakoff, G.
1968.
Instrumental adverbs and the concept of
deep structure.
Foundations of Language 4.4-29.
Lasnik, H., and R. Fiengo.
deletion.
Oehrle, R.
1973.
Complement object
Unpublished paper, M.I.T.
1974.
Some remarks on the painting of Rembrandt.
Unpublished paper, M.I.T.
Ross, J.R.
1967.
Constraints on variables in syntax.
Unpublished doctoral dissertation, M.I.T.
Wasow, T., and T. Roeper.
1971.
On the subject of gerunds.
Unpublished paper, M.I.T.
Williams, E.
1971.
paper, M.I.T.
Small clauses in English.
Unpublished
.
-
~-
.
I.....!
,
.
.
I
I
I
- . I , .
217
BIOGRAPHICAL NOTE
The author was born in Brooklyn, New York, on April 28,
1948.
In 1958, he moved with his family to a small town on
Long Island.
He did his undergraduate work in Spanish and
Russian literature at the State University of New York at
Stony Brook and received his bachelor's degree in June of
1969.
From the fall of 1969 to the present time, he has
been a graduate:student
in linguistics at M.I.T.
Starting
in the fall of 1974, he will spend a year as Visiting
Lecturer in linguistics at Princeton University.
-
--